15_car_crash/15_car_crash.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4256597f",
   "metadata": {},
   "source": [
    "`Описание проекта:` нужно создать систему, которая могла бы оценить риск ДТП по выбранному маршруту движения.\n",
    "\n",
    "Под риском понимается вероятность ДТП с любым повреждением транспортного средства. Как только водитель забронировал автомобиль, сел за руль и выбрал маршрут, система должна оценить уровень риска. Если уровень риска высок, водитель увидит предупреждение и рекомендации по маршруту."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "28615209",
   "metadata": {},
   "source": [
    "Импортируем библиотеки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "733469e4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting scikit-learn==1.1.2\n",
      "  Downloading scikit_learn-1.1.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (30.8 MB)\n",
      "\u001b[K     |████████████████████████████████| 30.8 MB 2.5 MB/s eta 0:00:01█████████             | 18.3 MB 2.5 MB/s eta 0:00:05\n",
      "\u001b[?25hRequirement already satisfied: numpy>=1.17.3 in /opt/conda/lib/python3.9/site-packages (from scikit-learn==1.1.2) (1.21.1)\n",
      "Requirement already satisfied: joblib>=1.0.0 in /opt/conda/lib/python3.9/site-packages (from scikit-learn==1.1.2) (1.1.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/lib/python3.9/site-packages (from scikit-learn==1.1.2) (3.1.0)\n",
      "Requirement already satisfied: scipy>=1.3.2 in /opt/conda/lib/python3.9/site-packages (from scikit-learn==1.1.2) (1.9.1)\n",
      "Installing collected packages: scikit-learn\n",
      "  Attempting uninstall: scikit-learn\n",
      "    Found existing installation: scikit-learn 0.24.1\n",
      "    Uninstalling scikit-learn-0.24.1:\n",
      "      Successfully uninstalled scikit-learn-0.24.1\n",
      "Successfully installed scikit-learn-1.1.2\n"
     ]
    }
   ],
   "source": [
    "!pip install scikit-learn==1.1.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5f7cb9bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from scipy.stats import loguniform\n",
    "from sqlalchemy import create_engine \n",
    "import seaborn as sns\n",
    "pd.set_option('display.max_columns', 500)\n",
    "\n",
    "from lightgbm import LGBMClassifier\n",
    "from catboost import CatBoostClassifier\n",
    "import sqlalchemy\n",
    "import sklearn\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "from sklearn.metrics import recall_score\n",
    "from sklearn.metrics import precision_score\n",
    "%matplotlib inline\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.metrics import  f1_score, confusion_matrix, make_scorer\n",
    "from sklearn.metrics import log_loss\n",
    "from sklearn.preprocessing import StandardScaler, OneHotEncoder, OrdinalEncoder\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from multiprocessing import Pool"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "5782e7e7",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-05-05T04:58:05.692448Z",
     "start_time": "2022-05-05T04:58:05.688565Z"
    }
   },
   "source": [
    "## Подключитесь к базе. Загрузите таблицы sql"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "89e6d061",
   "metadata": {},
   "source": [
    "\n",
    "Подключаемся к базе данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8e1f2d66",
   "metadata": {},
   "outputs": [],
   "source": [
    "db_config = {\n",
    "'user': 'praktikum_student',\n",
    "'pwd': 'Sdf4$2;d-d30pp',\n",
    "'host': 'rc1b-wcoijxj3yxfsf3fs.mdb.yandexcloud.net',\n",
    "'port': 6432, # порт подключения,\n",
    "'db': 'data-science-vehicle-db'\n",
    "} \n",
    "\n",
    "\n",
    "connection_string = 'postgresql://{}:{}@{}:{}/{}'.format(\n",
    "    db_config['user'],\n",
    "    db_config['pwd'],\n",
    "    db_config['host'],\n",
    "    db_config['port'],\n",
    "    db_config['db'],\n",
    ")\n",
    "\n",
    "\n",
    "con = create_engine(connection_string) "
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a82331f8",
   "metadata": {},
   "source": [
    "## Проведите первичное исследование таблиц"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "799f3eaa",
   "metadata": {},
   "source": [
    "Создадим функции для поиска пропущенных значений и дубликатов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "370cf0d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "def miss(data):\n",
    "    return data.isna().sum()\n",
    "\n",
    "def dup(data):\n",
    "    return data.duplicated().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "28b859a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "query = '''\n",
    "SELECT *\n",
    "FROM collisions\n",
    "where collision_date > '2012-01-01' and COLLISION_DAMAGE != 'scratch'\n",
    "'''\n",
    "df_collisions = pd.read_sql_query(query, con) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "19fb86c0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(case_id                            0\n",
       " county_city_location               0\n",
       " county_location                    0\n",
       " distance                           0\n",
       " direction                      23989\n",
       " intersection                     714\n",
       " weather_1                        412\n",
       " location_type                  69147\n",
       " collision_damage                   0\n",
       " party_count                        0\n",
       " primary_collision_factor         519\n",
       " pcf_violation_category          2335\n",
       " type_of_collision               1045\n",
       " motor_vehicle_involved_with      591\n",
       " road_surface                     802\n",
       " road_condition_1                 716\n",
       " lighting                         494\n",
       " control_device                   621\n",
       " collision_date                     0\n",
       " collision_time                  1306\n",
       " dtype: int64,\n",
       " 0)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "miss(df_collisions), dup(df_collisions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e2df95b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "query = '''\n",
    "SELECT *\n",
    "FROM parties\n",
    "where party_type = 'car'\n",
    "'''\n",
    "\n",
    "parties = pd.read_sql_query(query, con) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "03dd3e37",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(id                           0\n",
       " case_id                      0\n",
       " party_number                 0\n",
       " party_type                   0\n",
       " at_fault                     0\n",
       " insurance_premium       219689\n",
       " party_sobriety           44207\n",
       " party_drug_physical    2191243\n",
       " cellphone_in_use        374750\n",
       " dtype: int64,\n",
       " 0)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "miss(parties), dup(parties)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "45c90ef0",
   "metadata": {},
   "outputs": [],
   "source": [
    "query = '''\n",
    "SELECT *\n",
    "FROM vehicles\n",
    "'''\n",
    "vehicles = pd.read_sql_query(query, con) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "8cfa5b1d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(id                          0\n",
       " case_id                     0\n",
       " party_number                0\n",
       " vehicle_type                0\n",
       " vehicle_transmission    23659\n",
       " vehicle_age             24582\n",
       " dtype: int64,\n",
       " 0)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "miss(vehicles), dup(vehicles)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "80f8de8c",
   "metadata": {},
   "source": [
    "##  Проведите статистический анализ факторов ДТП"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7b7642ca",
   "metadata": {},
   "source": [
    "Проверим когда происходит наибольшее кол-во аварий по таблице collisions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "96f68224",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "query = '''SELECT extract(year from collision_date) AS year,\n",
    "           extract(month from collision_date) AS month,\n",
    "           count(case_id)\n",
    "FROM collisions where COLLISION_DAMAGE != 'scratch'\n",
    "group by extract(year from collision_date) , extract(month from collision_date)'''\n",
    "df_collisions = pd.read_sql_query(query, con)\n",
    "\n",
    "df_collisions['year'] = df_collisions['year'].astype(int).astype(str)\n",
    "df_collisions['month'] = df_collisions['month'].astype(int).astype(str)\n",
    "\n",
    "df_collisions['all'] = df_collisions['year'] + '-' + df_collisions['month']\n",
    "df_collisions\n",
    "\n",
    "df_collisions[['all', 'count']].plot(x = 'all',y = 'count', figsize = (12,6), kind = 'bar')\n",
    "plt.xticks(rotation=55) \n",
    "plt.xlabel('период')\n",
    "plt.ylabel('кол-во случаев ДТП')\n",
    "plt.title('Анализ случаев ДТП за период')\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "8210e9ae",
   "metadata": {},
   "source": [
    "`Вывод:` анализ ДТП не выявляет ярко выраженной тенденции, однако, можно заметить что общее количество ДТП имеет небольшую тенденцию к снижению в начале года (Январь). возможно, это связано с новогодними праздниками и тем что авто используется менее интенсивно чем в другие месяцы.\n",
    "\n",
    "Рост ДТП наблюдается в сентябре, скорее всего это связано с ухудшением погоды и началом осенних дождей.\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "8cd70fd3",
   "metadata": {},
   "source": [
    "Чтобы выяснить реальные причины, разобъём задачу на состовляющие, разделим задачу на подзадачи и передадим коллегам."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4abc832b",
   "metadata": {},
   "source": [
    "Примеры задач для коллег:"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "24961c66",
   "metadata": {},
   "source": [
    "\n",
    "- Изучить зависимость ДТП от состояния дорожного полотна\n",
    "- Изучить зависимость ДТП от состояния освящённости дороги\n",
    "- Изучить зависимость ДТП при нахождении на перекрёстке дорог\n",
    "- Изучить есть зависимость ДТП от уровня опасности участка дорого\n",
    "- Изучить зависимость ДТП от погодных условий\n",
    "- Изучить зависимость ДТП от использования громкой связи за рулём\n",
    "- Изучить зависимость ДТП от времени суток"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7e18a341",
   "metadata": {},
   "source": [
    "Выберем для анализа две задачи и проанализируем их:\n",
    "\n",
    "- Зависимость ДТП от погодных условий\n",
    "- Зависимость ДТП от состояния освящённости дороги"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "71474662",
   "metadata": {},
   "source": [
    "#### Задача 1: - Зависимость ДТП от погодных условий"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4fa8e8bf",
   "metadata": {},
   "source": [
    "- выполнить запрос к базе данных SQL\n",
    "- построить графическое отображение\n",
    "- сделать вывод о проделанной работе"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "6292650b",
   "metadata": {},
   "source": [
    "Запрос к базе данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f2d83226",
   "metadata": {},
   "outputs": [],
   "source": [
    "query = '''\n",
    "SELECT collisions.weather_1 , parties.at_fault\n",
    "FROM collisions\n",
    "LEFT JOIN parties on parties.case_id = collisions.case_id\n",
    "'''\n",
    "order_1 = pd.read_sql_query(query, con) "
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a0a9a427",
   "metadata": {},
   "source": [
    "График"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "db7de4f6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZcAAAEFCAYAAAAvyKpFAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAZMElEQVR4nO3deZSldX3n8fdHEBQRUJo4iJA22i4sitIuGCTuC0bREQ0Td4mMK3EimTExxjYTz8jRiR4TlRBFgUSBcMT0iMqoqCyydSM0W8AeQGlcAIFWIOzf+eN5yr42VV23up67Vb1f59Tpp577u7/7/VFNffvZvt9UFZIkdekBow5AkrTwmFwkSZ0zuUiSOmdykSR1zuQiSerclqMOYFwsWbKkli5dOuowJGlirF69+saq2mm610wuraVLl7Jq1apRhyFJEyPJj2d6zdNikqTOmVwkSZ0zuUiSOjcWySXJ15PsMIfxS5NcMsNrr0lyaZL7kizvLEhJUt/GIrlU1QFVdUtH010C/Gfg9I7mkyTN0VCSS5I/T3JYu/2JJKe1289L8i9JrkmypD0iuTzJP7VHH/83yYPbsfskuSjJRcC7Zvqsqrq8qq4YxrokSdMb1pHLGcCz2+3lwLZJHtju2/gIYxnw6araA7gFeHW7/wvAe6rqyV0FleTQJKuSrLrhhhu6mlaSFr1hJZfVwD5JtgPuBM6mSTLPpkk8va6uqgt73re0vR6zQ1VNJaLjugiqqo6qquVVtXynnaZ9DkiStBmGklyq6m7gauDNwA9oEspzgccCl280/M6e7XuZ5UHPJF9IcmGSr3cWsCRpXob5hP4ZwOHAW4GLgb8DVldVJdnkG6vqliS3JNmvqs4EXtfz2lsGGLMkaTMM826xM4CdgbOr6hfAHdz/lNimvAX4dJILgRmzUZJXJVkH7AuckuTUzQ9ZkrQ5YpvjxvLly8vaYpLUvySrq2ra5wnH4jkXSdLCYnKRJHXO5CJJ6pzJRZLUOZOLJKlzJhdJUufGIrl0XHL/4Um+leRH7Z8P6yxQSVJfxiK5dFxy//3Ad6pqGfCd9ntJ0hAtuJL7wIHAMe32McArB7cySdJ0FmLJ/UdU1c/a7Z8Dj5hn7JKkOVrQJferqW0zY30b+7lI0mAsxJL7v0iyc/vazsD1m4jLfi6SNADDrop8OM1psDOAtwM/rD4qZ7YX+29Jsl+767dK7lfV3lV1QLtrJfCmdvtNwL91E74kqV8LruQ+8FHghUl+BLyg/V6SNESW3G9Zcl+S5saS+5KkoTK5SJI6Z3KRJHXO5CJJ6pzJRZLUOZOLJKlzJhdJUueGnlyS7JDknT3fPyfJ1zqcf/8kFyS5J8lBXc0rSerfKI5cdgDeOdugfiXZuPbYT2hqmH2pq8+QJM3NwJNLkj9Lckn79V6aciyPaYtNfqwdtm2Sk5L8e9vfJe1790ny/SSrk5zaU5Dye0k+mWQV8Ke9n1dV11TVGuC+Qa9NkjS9TVYcnq8k+9DUBHsGTT2wc4HXA3tW1d7tmOcATwH2AH4KnAX8fpJzgb8HDqyqG5L8EfAR4K3t9FvNVHZgDvEdChwKsNtuu81nKklSj4EmF2A/4OSqug0gyVfY0DSs13lVta4dcyGwlKZR2J7At9oDmS2An/W854T5BldVRwFHQVNbbL7zSZIag04u/Zquh0uAS6tq3xneM5WwPgK8DGDqaEiSNFqDvuZyBvDKJNskeQjwKprTXg/t471XADsl2RcgyQOT7LHxoKr6QNvPZe8O45YkzcNAk0tVXQB8ETiP5nrL56pqNXBWe4H/Y5t4713AQcARSS4CLgSeNdtnJnlaknXAa4B/THLpvBciSZoT+7m07OciSXNjPxdJ0lCZXCRJnTO5SJI6Z3KRJHXO5CJJ6pzJRZLUOZOLJKlzC7Gfy9ZJTkiyNsm5SZZ2NbckqT8LsZ/LIcDNVfVY4BPAEV19liSpPwuunwtwIHBMu30S8Pyp+SRJw7EQ+7nsAlwLUFX3JFkP7AjcOE189nORpAGwn4v9XCSpcwuxn8t1wK7AuvZ6zPbALwcTtiRpOguxn8tK4E3t9kHAaWXpZ0kaqoEeuVTVBUm+SNPPBdp+LknOSnIJ8A3glBnee1eSg4BPJdm+jfWTwGz9WT4PHJdkLXATcPD8VyJJmgv7ubTs5yJJc2M/F0nSUJlcJEmdM7lIkjpncpEkdc7kIknqnMlFktS5sUguSZYn+VSH870kyRVt2f33dzWvJKk/Y1H+papWAZ08ZJJkC+DTwAuBdcD5SVZW1WVdzC9Jmt2sRy5JHpLklCQXtWXz/yjJNUk+nOSCJBcneUI79uFJvppkTZJzkjyp3X9x2yQsSX6Z5I3t/mOTvLC3YViSFUmObsvqX5XksJ5YPtgekZyZ5MtJDp8m5KcDa6vqqqq6Cziepgy/JGlI+jkt9hLgp1X15KraE/hmu//Gqnoq8Flg6pf8h4EfVtWTgL8Ejm33nwX8Pk1Z/avYUBl5X+AH03zmE4AX0ySKD7V1xZ4GvBp4MvBSYNqnQukpud9a1+67nySHJlmVZNUNN9ww0/olSXPUT3K5GHhhkiOSPLuq1rf7v9L+uZqmRD40JfaPA6iq04Adk2xHU8By//brs8BeSXah6Rh52zSfeUpV3VlVNwLXA4+gSU7/VlV3VNWvgf8zx7XeT1UdVVXLq2r5TjvtNN/pJEmtWZNLVV0JPJUmyfxtkr9uX5oqkz9VIn9TTqc5Wnk28D3gBpqKxWfMMH66EvzTSrJr29XywiRvZ0PJ/SmPavdJkoakn2sujwRur6p/Bj5Gk2hmcgbwuvZ9z6E5dfarqroWWAIsq6qrgDNpTqWdPodYzwJenuRBSbYF/hCgqq6dKrlfVUcC5wPLkjw6yVY0VZFXzuFzJEnz1M/dYnsBH0tyH3A38A6a3vTTWQEcnWQNcDsb+qpA0+J4i3b7DOB/0SSZvlTV+UlWAmuAX9AcSa2fZtw9Sd4NnNp+3tFVNVuZfklShyaq5H6Sbavq1iTb0Bz1HFpVF3QxtyX3JWluNlVyfyyec5mDo5LsDjwIOKarxCJJ6tZEJZeq+uNRxyBJmt1YlH+RJC0sJhdJUudMLpKkzplcJEmdM7lIkjo3FsllAP1cjk5yfZJLuppTktS/sUguVbWqqg6bfWTfvkhTzVmSNAILsZ8LVXU6cFMH/30kSZthIfZz6Zv9XCRpMOznYj8XSercQuznIkkasYXYz0WSNGL9nBbbCzgvyYXAh4C/3cTYFcA+bT+Xj3L/fi5Xtttn0PS1n1M/F5qmX2uAbzBDPxeAJF8GzgYen2RdkkP6/RxJ0vzZz6VlPxdJmhv7uUiShmqikov9XCRpMozFE/qSpIXF5CJJ6pzJRZLUOZNL6+Lrpr2rWZK0GcYmuSQ5LMnlSf5lnvMkyaeSrG0LaG7qoU9J0gCM091i7wReUFXr5jnPS4Fl7dczaGqZPWOec0qS5mAsjlySHAn8HvCNJO+boWz/Tkm+leTSJJ9L8uMkS6aZ7kDg2GqcA+yQZOchLkeSFr2xSC5V9Xbgp8BzaSosT1e2/0PAaVW1B3ASsNsM0+0CXNvz/bp23/30lty/93avuUhSV8YiuWxkprL9+wHHt/u/Cdw83w/qLbm/xTbbz3c6SVJrHJPLnCR5V0/J/UcC1wG79gx5VLtPkjQk45hcpi3bT1Ny/7Xt/hcBDwOoqk/3lNz/KU3l5De2d409E1hfVT8b/jIkafEap7vFpqwAjm7L9t/OhrL9Hwa+nOQNNOX0fw78epr3fx04AFjbvv8t/XzoXrt4WkySujI2yaWqlvZ8+8pphqwHXlxV9yTZF3haVd258aBqegi8ayBBSpL6MjbJpQ+7AScmeQBwF/C2EccjSZrBxCSXqvoR8JRRxyFJmt04XtCXJE04k4skqXMmF0lS50wukqTOmVxa9nORpO6MTXLpsJ/LE5KcneTOJId3FZ8kqX/jdCtyV/1cbgIOY/oHMSVJQzAWRy5d9nOpquur6nzg7iEvQ5LUGovk0nE/l77Zz0WSBmMskstG7OciSRNuHJPLnEzTz0WSNGLjmFzm289FkjRi43S32JQVzKOfS5L/BKwCtgPuS/JeYPc2Qc3Ifi6S1J2xSS4d9nP5OU1rY0nSiIxNcumD/VwkaUJMTHKxn4skTY5xvKAvSZpwJhdJUudMLpKkzk3MNZdBu/i69Sx9/ylzes81H33ZgKKRpMk28COXtsjk7rOMeXuSN3b4mX+RZG2SK5K8uKt5JUn96eTIJUmAVNV9G79WVX8y2/ur6sgu4mhj2R04GNgDeCTw7SSPq6p7u/oMSdKmbfaRS5Kl7ZHBscAlwOfbCsOXJvlwz7jvJVnebt+a5CNJLmrL6T+i3b9iqrFXO/6IJOcluTLJs9v92yQ5McllSU5Ocu7UvBs5EDi+qu6sqquBtcDTN3edkqS5m+9psWXAZ9oy+O+rquXAk4A/mOrDspGHAOdU1ZOB05n5Qcgtq+rpwHtpSu1D00zs5qraHfggsM8M790FuLbn+3Xtvvux5L4kDcZ8k8uPq+qcdvu1SS4AfkhzSmq66yx3AV9rt1fT9G6ZzlemGdNbcv8SYM18Am/nseS+JA3AfJPLbQBJHg0cDjy/bfJ1CvCgacbfXVXVbt/LzNd87uxjDO1nv6qn5P5y4Dpg154hj2r3SZKGpKu7xbajSTTr2+soL+1o3l69Jfd3B/YCqKqTe0rurwJWAgcn2bpNesuA8wYQjyRpBp3cLVZVFyX5IfDvNNc7zupi3o18BjgmyWXt51xKUyl541guTXIicBlwD/Cufu4U22uX7VnlcyuS1IlsOEs13pJsATywqu5I8hjg28Djq+quLuZfvnx5rVq1qoupJGlRSLK6vZHrfibpCf1tgO8meSAQ4J1dJRZJUrcmJrlU1a+BaTOkJGm8WLhSktQ5k4skqXMmF0lS50wukqTOTcwF/UHbnH4uYE8XSZrOguvnkmTHJN9tKzD/QxdzSpLmZsH1cwHuoKmavGf7JUkasgXXz6WqbquqM2mSjCRpBBZiP5e+2c9FkgbDfi72c5Gkzi3Efi6SpBFbiP1cJEkjtuD6uQAkuYYm4W2V5JXAi6rqsk1Nbj8XSeqO/Vxa9nORpLmxn4skaagmJrnYz0WSJoeFKyVJnTO5SJI6Z3KRJHXO5CJJ6txALugnWQHcWlUf72Cu7wGH9/uAZJL9gU/S1Dg7uKpO6ud9m9vPZdzYX0bSOFiIRy4/Ad4MfGnEcUjSotVJcknyxiRr2lL6x2302t5tef01ban8h7X7e0vxL2mfqifJg5Mcn+TyJCcDD273vzXJJ3vmfVuST2wcS1VdU1VrgPv1lpEkDce8k0uSPYC/Ap7XltL/042GHAv8j7ag5cVsKKE/k3cAt1fVE9uxU6X1TwRe3j5ECfAW4Oh5xm7JfUkagC6OXJ4H/GtV3QhQVTdNvZBke2CHqvp+u+sYYP9Z5tsf+Od2rjW0pfWr6lbgNOAPkzyBphTMxfMJ3JL7kjQYo3xC/x42JLfpyvNP53PAX9IUrvwCQJKPAC8DqKq9uw1RkrQ5ujhyOQ14TZIdAZI8fOqFqloP3DzVqhh4AzB1FHMNG055HdQz3+nAH7dz7Ulz19fUfOcCu7avf7nd94GpkvsdrEWS1IF5H7lU1aXt0cP3k9xL04nymp4hbwKOTLINcBXNtRKAjwMnJjmUprnYlM8CX0hyOXA5TTfKXicCe1fVzdPFk+RpwMnAw2iu0Xy4bcO8SZbcl6TuTEzJ/SlJvgZ8oqq+0+W8ltyXpLnZVMn9iXnOJckOSa4E/qPrxCJJ6tYkldy/BXjcqOOQJM1uYo5cJEmTw+QiSeqcyUWS1DmTiySpcxNzQX/QFkrJ/SmW3pc0SgM5ckmyIsnhHc31m+rJfY7fOskJSdYmOTfJ0i7ikCT1byGeFjsEuLmqHgt8AjhixPFI0qKz4Pq5AAfSVF8GOAl4fpJ0sU5JUn8WYj+XXYBrAarqHmA9sOMMsdvPRZIGwH4u9nORpM6N8prL5vZzeTPNUctv+rkkuTDJhe2Y62jK8pNkS2B74JfdhCxJ6sdC7OeykqbM/9S8p9WklX6WpAm34Pq5AJ8HjkuyFrgJOLifddjPRZK6Yz+Xlv1cJGlu7OciSRqqiSn/Yj8XSZocE3PkIkmaHCYXSVLnTC6SpM6ZXCRJnRvZBf0kK4Bbq+rjHc+7I03ByqcBX6yqd/fzvoXWz0XS8NlHaYOJuVtsJkm2bAtUTrkD+CCwZ/slSRqyoZ0Wm6Us/2OSfDPJ6iRntIUpSfLytuHXD5N8O8kj2v0rkhyX5Czgt+aqqtuq6kyaJCNJGoGhHLn0lOV/VlXd2NYfO6xnyFHA26vqR0meAXyGptrymcAzq6qS/Anw34H3te/ZHdivqv5jGGuQJPVvWKfF7leWf6p/V5JtgWcB/9rT02vr9s9HASck2RnYCri6Z86V800sbV2zQwG22G6n+UwlSeoxDneLPQC4Zaqycfv1xPa1vwf+oar2Av4rv12a/zaAJK+aKrk/1dmyX/ZzkaTBGFZy2VRZ/l8BVyd5Tftakjy5fXl7mv4ssKGM/m+pqpN7kpKVJyVpDAzltFgfZflfB3w2yV8BDwSOBy4CVtCcLruZJkE9up/PS3INsB2wVZJXAi+qqss29R5L7ktSdyau5P6gWHJfkuZmQZTclyRNDpOLJKlzJhdJUudMLpKkzplcJEmdM7lIkjo38VWRu2LJfUmLzSBbBIzsyKWtbHz4gOb+iyRrk1yR5MWD+AxJ0swm/shl434uSXYHDgb2AB4JfDvJ46rq3lHFKEmLzYLr5wIcCBxfVXdW1dXAWuDpQ1iiJKm1EPu57AKc0/P9unbfdHFZcl+SBmBR93OpqqNoEhtb77zMImuS1JFxuBW5634u1wG79ox7FBvK9kuShmAh9nNZCRycZOskjwaWAecNZFWSpGktuH4u7WedCFwG3AO8q587xeznIkndsZ9Ly34ukjQ39nORJA2VyUWS1DlPi7WS/Bq4YtRxdGwJcOOog+iYa5oMC3FNsDDXNZ81/W5VTfuQ4MSXf+nQFTOdO5xUSVa5pvHnmibHQlzXoNbkaTFJUudMLpKkzplcNjhq1AEMgGuaDK5pcizEdQ1kTV7QlyR1ziMXSVLnTC6SpM4tquSS5CVt6+O1Sd4/zetbJzmhff3cJEtHEOac9LGmP0tyWduo7TtJfncUcc7VbOvqGffqJNVWxB5r/awpyWvbn9elSb407Bjnqo+/f7sl+W7b8G9NkgNGEedcJDk6yfVJLpnh9ST5VLvmNUmeOuwY56qPNb2uXcvFSX7QUzx481XVovgCtgD+H/B7NL1hLgJ232jMO4Ej2+2DgRNGHXcHa3ousE27/Y5xX1O/62rHPRQ4naY53PJRx93Bz2oZTVHXh7Xf/86o4+5gTUcB72i3dweuGXXcfaxrf+CpwCUzvH4A8A0gwDOBc0cdcwdrelbP37uXdrGmxXTk8nRgbVVdVVV30VRePnCjMQcCx7TbJwHPT08HszE065qq6rtVdXv77Tk0/W3GXT8/K4D/CRwB3DHM4DZTP2t6G/DpqroZoKquH3KMc9XPmgrYrt3eHvjpEOPbLFV1OnDTJoYcCBxbjXOAHdqGhmNrtjVV1Q+m/t7R0e+JxZRcdgGu7fl+uvbHvxlTVfcA64EdhxLd5ulnTb0OofkX17ibdV3tqYhdq+qUYQY2D/38rB4HPC7JWUnOSfKSoUW3efpZ0wrg9UnWAV8H3jOc0AZqrv/fTZpOfk9Y/mWRSPJ6YDnwB6OOZb6SPAD4O+DNIw6la1vSnBp7Ds2/HE9PsldV3TLKoObpvwBfrKr/nWRf4Lgke1bVfaMOTPeX5Lk0yWW/+c61mI5c+ml//JsxSbakOYz/5VCi2zx9tXRO8gLgA8ArqurOIcU2H7Ot66HAnsD3klxDc9575Zhf1O/nZ7UOWFlVd1fV1cCVNMlmXPWzpkOAEwGq6myaVuVLhhLd4CzIVupJngR8Djiwqub9e28xJZfzgWVJHp1kK5oL9is3GrOSDe2UDwJOq/YK15iadU1JngL8I01iGfdz+FM2ua6qWl9VS6pqaVUtpTlH/Ipq2lyPq37+/n2V5qiFJEtoTpNdNcQY56qfNf0EeD5AkifSJJcbhhpl91YCb2zvGnsmsL6qfjbqoOYjyW7AV4A3VNWVXcy5aE6LVdU9Sd4NnEpzl8vR1bRE/htgVVWtBD5Pc9i+lubi18Gji3h2fa7pY8C2NO2iAX5SVa8YWdB96HNdE6XPNZ0KvCjJZcC9wJ938S/IQelzTe8D/inJf6O5uP/mMf8HG0m+TJPkl7TXij5E036dqjqS5trRAcBa4HbgLaOJtH99rOmvaa4vf6b9PXFPzbNSsuVfJEmdW0ynxSRJQ2JykSR1zuQiSeqcyUWS1DmTiyQtQrMVs5xm/JyKqnq3mCQtQkn2B26lqZO25yxjl9E8DPu8qro5ye/M9tycRy6StAhNV8wyyWOSfDPJ6iRnJHlC+9Kci6qaXCRJU44C3lNV+wCHA59p98+5qOqieUJfkjSzJNvS9HWZquYBsHX755yLqppcJEnQnMm6par2nua1dTQNxO4Grk4yVVT1/E1NJkla5KrqVzSJ4zXwm3bOU+2Ov8oci6qaXCRpEWqLWZ4NPD7JuiSHAK8DDklyEXApGzqLngr8si2q+l36KKrqrciSpM555CJJ6pzJRZLUOZOLJKlzJhdJUudMLpKkzplcJEmdM7lIkjr3/wHztmpBycDG8AAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "order_1['at_fault'] = order_1['at_fault'].astype(str)\n",
    "order_1['all'] = order_1['weather_1'] + '-' + order_1['at_fault']\n",
    "order_1['all'].value_counts().plot(kind = 'barh')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "db24321a",
   "metadata": {},
   "source": [
    "`Вывод:` из графика выше следует что за исключением параметра 'ясная погода' сколько нибудь значительной разницы между присутствие или отсутствием погодного явления не надлюдается.\n",
    "\n",
    "ДТП проиходят одинаково часто в дождь или при его отсутствии.\n",
    "\n",
    "Папаметр clear - ясня погода показывает незначительную связь, на грани погрешности."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "b4ab8a1c",
   "metadata": {},
   "source": [
    "#### Задача 2: Зависимость кол-во ДТП от состояния освящённости дороги"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "52d9e7df",
   "metadata": {},
   "source": [
    "- выполнить запрос к базе данных SQL\n",
    "- построить графическое отображение\n",
    "- сделать вывод о проделанной работе"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7fc4ef9e",
   "metadata": {},
   "source": [
    "Запрос к базе данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "1a8dbf8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "query = '''\n",
    "SELECT collisions.LIGHTING , parties.at_fault\n",
    "FROM collisions\n",
    "LEFT JOIN parties on parties.case_id = collisions.case_id\n",
    "'''\n",
    "order_2 = pd.read_sql_query(query, con) "
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "00421318",
   "metadata": {},
   "source": [
    "График"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "8bd53591",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "order_2['at_fault'] = order_2['at_fault'].astype(str)\n",
    "order_2['all'] = order_2['lighting'] + '-' + order_2['at_fault']\n",
    "order_2['all'].value_counts().plot(kind = 'barh')\n",
    "ax = order_2['all'].value_counts().plot(kind='barh')\n",
    "\n",
    "# название таблицы\n",
    "ax.set_title(\"Частота по всем категориям\")\n",
    "\n",
    "# ось Х\n",
    "ax.set_xlabel(\"Частота\")\n",
    "\n",
    "# ось у\n",
    "ax.set_ylabel(\"Все факторы\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "1bdc8126",
   "metadata": {},
   "source": [
    "`Вывод:` не прослеживается корреляции отсещённости и ДТП, за исключением наличия / отсутствия дневного света, но эти показатели очень близки и поэтому, ненадёжны. Буди считать что явной взаимосвязи нет."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7dd3aac5",
   "metadata": {},
   "source": [
    "## Создайте модель для оценки водительского риска"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "4c856963",
   "metadata": {},
   "source": [
    "Запрос к базе данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "4ac2fc0f",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "query = '''\n",
    "SELECT *\n",
    "FROM collisions as c\n",
    "left join parties as p on p.case_id = c.case_id\n",
    "left join vehicles as v on v.case_id = c.case_id\n",
    "where extract(year from collision_date) = 2012 and COLLISION_DAMAGE != 'scratch' and party_type = 'car'\n",
    "'''\n",
    "union_df = pd.read_sql_query(query, con) "
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7b729ada",
   "metadata": {},
   "source": [
    "Соединяем все три таблицы чтобы работать с одной таблицей"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "8aceae0d",
   "metadata": {},
   "source": [
    "Изучаем и выбираем факторы которые могуть повлиять на вероятность ДТП"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "3f10e487",
   "metadata": {},
   "outputs": [],
   "source": [
    "columns= ['case_id',  # после объединения колонка не нужна -- удаляем\n",
    "        'county_city_location', # числовое описание района -- оставляем\n",
    "        'county_location',  # дублирует `county_city_location` -- удаляем\n",
    "        'distance', # протяженность пути -- оставляем\n",
    "       'direction', # неточная формулировка (стороны света) -- убираем\n",
    "       'intersection', # наличие перекрёстка в месте ДТП -- оставляем\n",
    "       'weather_1', # погодняе условия -- оставляем\n",
    "       'location_type',# тип дороги -- оставляем\n",
    "       'collision_damage', # повреждения не интересуют, важен факт ДТП -- удаляем\n",
    "       'party_count', # Обьект столковения нас не интересует -- удаляем\n",
    "       'primary_collision_factor',# какие правила ДТП нарушены -- оставляем\n",
    "       'pcf_violation_category',# наличие алкоголья -- оставляем\n",
    "       'type_of_collision',# тип ДТП не имеет значения -- удаляем\n",
    "       'motor_vehicle_involved_with',# побочные участники дтп вносят шум -- удаляем\n",
    "       'road_surface', # дорожное покрытие -- оставляем\n",
    "       'road_condition_1', # Состояние дорожного полотна - важно, оставляем\n",
    "       'lighting',  # освещение дорожного полотна -- оставляем\n",
    "       'control_device', # нет описания -- оставим на всякий случай\n",
    "        'collision_date'# дата ДТП -- оставляем\n",
    "        'collision_time',# время ДТП -- оставляем\n",
    "         'id',# не несёт полезной инфы -- удаляем \n",
    "       'case_id',# не несёт полезной инфы -- удаляем \n",
    "        'party_number',#все порядковые колонки с удаляем\n",
    "        'party_type',# удаляем так как мы уже выбрали ранее все автомобили\n",
    "         'at_fault',# целевой параметр -- оставляем\n",
    "       'insurance_premium',# нет описания опасности поетому -- удаляем \n",
    "        'party_sobriety', # нет данных -- удаляем       \n",
    "        'party_drug_physical',# состояние участника на тесте нам не известно -- удаляем    \n",
    "       'cellphone_in_use',# использование телефона во время вождения понижает внимательность -- оставляем\n",
    "        'id',# не несёт полезной инфы -- удаляем \n",
    "       'case_id',# не несёт полезной инфы -- удаляем \n",
    "        'party_number',# не несёт полезной инфы -- удаляем \n",
    "        'vehicle_type',#тип кузова -- может повлиять на ДТП, тк от типа кузова может зависеть поведение авто и его управляемость -- оставим\n",
    "       'vehicle_transmission',#  трансмиссия -- оставим, возможно механника как более сложный в управлении тип влияет на ДТП\n",
    "         'vehicle_age']# возраст авто -- оставляем"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "5695ff55",
   "metadata": {},
   "source": [
    "На основании анализа выше - удалим ненужные признаки и сохраним в переменную union_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "d4bd348d",
   "metadata": {},
   "outputs": [],
   "source": [
    "union_df = union_df[['county_city_location','pcf_violation_category','primary_collision_factor','party_sobriety','distance','intersection','weather_1','road_surface','at_fault','road_condition_1','lighting','control_device','collision_date','collision_time','cellphone_in_use','vehicle_age', 'vehicle_type', 'vehicle_transmission']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "c45e7723",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>county_city_location</th>\n",
       "      <th>pcf_violation_category</th>\n",
       "      <th>primary_collision_factor</th>\n",
       "      <th>party_sobriety</th>\n",
       "      <th>distance</th>\n",
       "      <th>intersection</th>\n",
       "      <th>weather_1</th>\n",
       "      <th>road_surface</th>\n",
       "      <th>at_fault</th>\n",
       "      <th>road_condition_1</th>\n",
       "      <th>lighting</th>\n",
       "      <th>control_device</th>\n",
       "      <th>collision_date</th>\n",
       "      <th>collision_time</th>\n",
       "      <th>cellphone_in_use</th>\n",
       "      <th>vehicle_age</th>\n",
       "      <th>vehicle_type</th>\n",
       "      <th>vehicle_transmission</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3000</td>\n",
       "      <td>unsafe lane change</td>\n",
       "      <td>vehicle code violation</td>\n",
       "      <td>had not been drinking</td>\n",
       "      <td>900.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>clear</td>\n",
       "      <td>dry</td>\n",
       "      <td>0</td>\n",
       "      <td>normal</td>\n",
       "      <td>daylight</td>\n",
       "      <td>none</td>\n",
       "      <td>2012-01-26</td>\n",
       "      <td>08:50:00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0900</td>\n",
       "      <td>unknown</td>\n",
       "      <td>unknown</td>\n",
       "      <td>had not been drinking</td>\n",
       "      <td>200.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>clear</td>\n",
       "      <td>dry</td>\n",
       "      <td>0</td>\n",
       "      <td>normal</td>\n",
       "      <td>daylight</td>\n",
       "      <td>functioning</td>\n",
       "      <td>2012-01-15</td>\n",
       "      <td>16:15:00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>sedan</td>\n",
       "      <td>manual</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3026</td>\n",
       "      <td>unsafe lane change</td>\n",
       "      <td>vehicle code violation</td>\n",
       "      <td>had not been drinking</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>clear</td>\n",
       "      <td>dry</td>\n",
       "      <td>0</td>\n",
       "      <td>normal</td>\n",
       "      <td>daylight</td>\n",
       "      <td>none</td>\n",
       "      <td>2012-01-03</td>\n",
       "      <td>09:16:00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3801</td>\n",
       "      <td>speeding</td>\n",
       "      <td>vehicle code violation</td>\n",
       "      <td>impairment unknown</td>\n",
       "      <td>22.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>other</td>\n",
       "      <td>None</td>\n",
       "      <td>1</td>\n",
       "      <td>normal</td>\n",
       "      <td>dark with street lights</td>\n",
       "      <td>none</td>\n",
       "      <td>2012-01-22</td>\n",
       "      <td>05:55:00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1941</td>\n",
       "      <td>improper turning</td>\n",
       "      <td>vehicle code violation</td>\n",
       "      <td>had not been drinking</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>clear</td>\n",
       "      <td>dry</td>\n",
       "      <td>1</td>\n",
       "      <td>normal</td>\n",
       "      <td>daylight</td>\n",
       "      <td>functioning</td>\n",
       "      <td>2012-01-18</td>\n",
       "      <td>07:22:00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>hatchback</td>\n",
       "      <td>auto</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  county_city_location pcf_violation_category primary_collision_factor  \\\n",
       "0                 3000     unsafe lane change   vehicle code violation   \n",
       "1                 0900                unknown                  unknown   \n",
       "2                 3026     unsafe lane change   vehicle code violation   \n",
       "3                 3801               speeding   vehicle code violation   \n",
       "4                 1941       improper turning   vehicle code violation   \n",
       "\n",
       "          party_sobriety  distance  intersection weather_1 road_surface  \\\n",
       "0  had not been drinking     900.0           0.0     clear          dry   \n",
       "1  had not been drinking     200.0           0.0     clear          dry   \n",
       "2  had not been drinking       1.0           0.0     clear          dry   \n",
       "3     impairment unknown      22.0           0.0     other         None   \n",
       "4  had not been drinking       0.0           1.0     clear          dry   \n",
       "\n",
       "   at_fault road_condition_1                 lighting control_device  \\\n",
       "0         0           normal                 daylight           none   \n",
       "1         0           normal                 daylight    functioning   \n",
       "2         0           normal                 daylight           none   \n",
       "3         1           normal  dark with street lights           none   \n",
       "4         1           normal                 daylight    functioning   \n",
       "\n",
       "  collision_date collision_time  cellphone_in_use  vehicle_age vehicle_type  \\\n",
       "0     2012-01-26       08:50:00               0.0          NaN         None   \n",
       "1     2012-01-15       16:15:00               0.0          4.0        sedan   \n",
       "2     2012-01-03       09:16:00               0.0          NaN         None   \n",
       "3     2012-01-22       05:55:00               NaN          NaN         None   \n",
       "4     2012-01-18       07:22:00               0.0          2.0    hatchback   \n",
       "\n",
       "  vehicle_transmission  \n",
       "0                 None  \n",
       "1               manual  \n",
       "2                 None  \n",
       "3                 None  \n",
       "4                 auto  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "union_df.head()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "d8d18dcf",
   "metadata": {},
   "source": [
    "Заполним пропуски"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "b732f3da",
   "metadata": {},
   "outputs": [],
   "source": [
    "union_df.loc[union_df['vehicle_age'].isna(),'vehicle_age'] = 0\n",
    "union_df.loc[union_df['vehicle_transmission'].isna(), 'vehicle_transmission' ] = 'none'\n",
    "union_df.loc[union_df['vehicle_type'].isna(), 'vehicle_type' ] = 'none'\n",
    "union_df.loc[union_df['cellphone_in_use'].isna(),'cellphone_in_use'] = 0\n",
    "union_df.loc[union_df['intersection'].isna(),'intersection'] = 0\n",
    "union_df.loc[union_df['control_device'].isna(),'control_device'] = 'none'\n",
    "union_df.loc[union_df['lighting'].isna(),'lighting'] = 'none'\n",
    "union_df.loc[union_df['weather_1'].isna(),'weather_1'] = 'other'\n",
    "union_df = union_df.dropna(subset=['road_surface','road_condition_1','pcf_violation_category','primary_collision_factor','party_sobriety']) # удаляем пропуски"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "e266a937",
   "metadata": {},
   "source": [
    "Проверим на выбросы расстояние от главной дороги 'distance'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "82fc302f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "union_df['distance'].plot()\n",
    "plt.ticklabel_format(style='plain', axis='y')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "f7959993",
   "metadata": {},
   "source": [
    "Проверим на выбросы расстояние от главной дороги 'vehicle_age'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "6d4e91fa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "union_df['vehicle_age'].plot()\n",
    "plt.ticklabel_format(style='plain', axis='y')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "aabce450",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0      112726\n",
       "3.0       20882\n",
       "4.0       14119\n",
       "2.0       12390\n",
       "5.0       11305\n",
       "6.0        8201\n",
       "7.0        8030\n",
       "8.0        7360\n",
       "9.0        5873\n",
       "1.0        5480\n",
       "10.0       4132\n",
       "11.0       2885\n",
       "12.0       1807\n",
       "13.0       1137\n",
       "14.0        583\n",
       "15.0         79\n",
       "16.0         20\n",
       "17.0          4\n",
       "161.0         4\n",
       "19.0          1\n",
       "Name: vehicle_age, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "union_df['vehicle_age'].value_counts()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "48572013",
   "metadata": {},
   "source": [
    "Удалим выбирающиеся значения большие 200 000 метров от главной дороги"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "10cac50c",
   "metadata": {},
   "outputs": [],
   "source": [
    "union_df = union_df.loc[union_df['distance'] < 200000]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "869a17f8",
   "metadata": {},
   "source": [
    "Удалим выбирающиеся значения возраста авто большие 19 лет."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c1578b28",
   "metadata": {},
   "outputs": [],
   "source": [
    "union_df = union_df.loc[union_df['vehicle_age'] <= 19]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7a8d42ee",
   "metadata": {},
   "source": [
    "Готовим данные перед разделением выборок"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "90cc870f",
   "metadata": {},
   "outputs": [],
   "source": [
    "categorical = union_df.select_dtypes(exclude=np.number).columns.to_list()\n",
    "numeric = union_df.select_dtypes(include=np.number).columns.to_list()\n",
    "numeric = ['distance', 'intersection', 'cellphone_in_use', 'vehicle_age']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "7e76ca54",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = union_df.drop(['at_fault'], axis=1)\n",
    "y = union_df['at_fault']"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "ab0bd939",
   "metadata": {},
   "source": [
    "Делим выборки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "b7c5c5e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Разделение на выборки\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, stratify=y, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "c6b9dd58",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Разделение на выборки\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, train_size=0.5, stratify=y_train, random_state=42)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "e1893786",
   "metadata": {},
   "source": [
    "Создаём пайплайны"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "7471b36e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# масштабирование числовых данных\n",
    "numeric_preproces = Pipeline([('scaler',StandardScaler())])\n",
    "\n",
    "# кодирование категориальных данных \n",
    "categorial_preproces = Pipeline([('ordinal', OrdinalEncoder(handle_unknown='use_encoded_value', unknown_value=-1))])\n",
    "\n",
    "# процесс подготовки данных для Случайного леса\n",
    "preprocessor_for_tree = ColumnTransformer([\n",
    "    (\n",
    "        'categorical', \n",
    "        categorial_preproces, \n",
    "        categorical\n",
    "    ),\n",
    "    (\n",
    "        'numerical', \n",
    "        numeric_preproces, \n",
    "        numeric\n",
    "    )\n",
    "])\n",
    "\n",
    "# подготовка данных для бустинга\n",
    "preprocessor_for_boost = ColumnTransformer([\n",
    "    (\n",
    "        'numerical', \n",
    "        numeric_preproces, \n",
    "        numeric\n",
    "    )\n",
    "])\n",
    "\n",
    "pipe_for_tree = Pipeline([\n",
    "    ('preprocessor', preprocessor_for_tree),\n",
    "    ('classifier', RandomForestClassifier(random_state=42))\n",
    "])\n",
    "\n",
    "pipe_for_catboost = Pipeline([\n",
    "    ('preprocessor', preprocessor_for_boost),\n",
    "    ('classifier', CatBoostClassifier(random_state=42))\n",
    "])\n",
    "\n",
    "\n",
    "pipe_for_lgbm = Pipeline([\n",
    "    ('preprocessor', preprocessor_for_boost),\n",
    "    ('classifier', LGBMClassifier(random_state=42))\n",
    "])"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "feae2462",
   "metadata": {},
   "source": [
    "Проверка на дисбаланс классов (0 и 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "f2e5e28b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Распределение классов:\n",
      "1    109578\n",
      "0    107428\n",
      "Name: at_fault, dtype: int64\n",
      "\n",
      "Соотношение классов:\n",
      "1    0.504954\n",
      "0    0.495046\n",
      "Name: at_fault, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# Провидим анализ дисбаланса классов, считаем кол-во 0 и 1 и их соотношение\n",
    "target_counts = union_df['at_fault'].value_counts()\n",
    "target_ratio = target_counts / len(union_df)\n",
    "\n",
    "print(\"Распределение классов:\")\n",
    "print(target_counts)\n",
    "print(\"\\nСоотношение классов:\")\n",
    "print(target_ratio)\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "3b44db3d",
   "metadata": {},
   "source": [
    "`Вывод:` дисбаланса класов нет"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "00ec4b56",
   "metadata": {},
   "source": [
    "`Выбор метрики:` обоснование выбора метрики F1 для системы оценки риска ДТП по выбранному маршруту движения можно представить следующим образом:\n",
    "\n",
    "Учет баланса между полнотой и точностью: Метрика F1 учитывает как полноту (recall), так и точность (precision) модели. В случае оценки риска ДТП, важно достичь баланса между выявлением действительно рисковых маршрутов (высокая полнота) и минимизацией ложно-положительных предупреждений (высокая точность).\n",
    "\n",
    "Подчеркивание важности правильного выявления рисковых маршрутов: В контексте безопасности дорожного движения, цель системы - предупреждать водителей о потенциальных опасностях на выбранных ими маршрутах. Ошибка в пропуске рискового маршрута (ложно-отрицательное предупреждение) может иметь серьезные последствия. Метрика F1 помогает оценить эффективность системы в правильном выявлении и предупреждении о рисковых маршрутах, уделяя внимание и полноте, и точности.\n",
    "\n",
    "\n",
    "Легкость интерпретации: Метрика F1 является широко используемой и легко интерпретируемой метрикой. Ее значения лежат в диапазоне от 0 до 1, где 1 соответствует идеальному балансу полноты и точности, а 0 - плохим результатам. Таким образом, F1-метрика предоставляет простую и понятную оценку уровня риска для водителей, позволяя"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "889130c5",
   "metadata": {},
   "source": [
    "Параметры для RF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "fea2085f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train_scores_1:0.5884641173656728\n",
      "val_scores_1:0.5861726655565518\n"
     ]
    }
   ],
   "source": [
    "param = {\n",
    "    'classifier__n_estimators': [100, 200],\n",
    "    'classifier__max_depth': [5, 10]\n",
    "}\n",
    "grid_tree =  GridSearchCV(pipe_for_tree, param, cv = 2, scoring= 'f1')\n",
    "grid_tree.fit(X_train,y_train)\n",
    "\n",
    "predict_train_rf = grid_tree.predict(X_train)\n",
    "predict_test_rf = grid_tree.predict(X_val)\n",
    "\n",
    "train_scores_1 = grid_tree.best_score_\n",
    "val_scores_1 = f1_score(y_val, predict_test_rf)\n",
    "print(f'train_scores_1:{train_scores_1}')\n",
    "print(f'val_scores_1:{val_scores_1}')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a49eb8bd",
   "metadata": {},
   "source": [
    "Параметры для LGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "98756182",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train_scores_2:0.5306453139611587\n",
      "val_scores_1:0.5349204226293968\n"
     ]
    }
   ],
   "source": [
    "param = {\n",
    "    'classifier__n_estimators': [100, 200],\n",
    "    'classifier__max_depth': [5, 10]\n",
    "}\n",
    "grid_lgbm = GridSearchCV(pipe_for_lgbm, param, cv = 2, scoring= 'f1', verbose=False)\n",
    "grid_lgbm.fit(X_train, y_train)\n",
    "predict_train_lgbm = grid_lgbm.predict(X_train)\n",
    "predict_test_lgbm = grid_lgbm.predict(X_val)\n",
    "train_scores_2 = grid_lgbm.best_score_\n",
    "val_scores_2 = f1_score(y_val, predict_test_lgbm)\n",
    "print(f'train_scores_2:{train_scores_2}')\n",
    "print(f'val_scores_1:{val_scores_2}')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "6ad14471",
   "metadata": {},
   "source": [
    "Параметры для Catboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "361b4056",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.402147\n",
      "0:\tlearn: 0.6903771\ttotal: 52.2ms\tremaining: 5.17s\n",
      "1:\tlearn: 0.6891517\ttotal: 56.7ms\tremaining: 2.78s\n",
      "2:\tlearn: 0.6885510\ttotal: 59.9ms\tremaining: 1.94s\n",
      "3:\tlearn: 0.6880931\ttotal: 63.2ms\tremaining: 1.52s\n",
      "4:\tlearn: 0.6879430\ttotal: 67.7ms\tremaining: 1.29s\n",
      "5:\tlearn: 0.6878568\ttotal: 70.3ms\tremaining: 1.1s\n",
      "6:\tlearn: 0.6877677\ttotal: 73.5ms\tremaining: 976ms\n",
      "7:\tlearn: 0.6876592\ttotal: 77.2ms\tremaining: 888ms\n",
      "8:\tlearn: 0.6876471\ttotal: 80.4ms\tremaining: 813ms\n",
      "9:\tlearn: 0.6875244\ttotal: 83.7ms\tremaining: 753ms\n",
      "10:\tlearn: 0.6875078\ttotal: 92.4ms\tremaining: 748ms\n",
      "11:\tlearn: 0.6874587\ttotal: 98.2ms\tremaining: 720ms\n",
      "12:\tlearn: 0.6873732\ttotal: 102ms\tremaining: 680ms\n",
      "13:\tlearn: 0.6872586\ttotal: 105ms\tremaining: 645ms\n",
      "14:\tlearn: 0.6872126\ttotal: 108ms\tremaining: 614ms\n",
      "15:\tlearn: 0.6870938\ttotal: 112ms\tremaining: 587ms\n",
      "16:\tlearn: 0.6870569\ttotal: 115ms\tremaining: 562ms\n",
      "17:\tlearn: 0.6870238\ttotal: 118ms\tremaining: 538ms\n",
      "18:\tlearn: 0.6869968\ttotal: 121ms\tremaining: 517ms\n",
      "19:\tlearn: 0.6869480\ttotal: 125ms\tremaining: 499ms\n",
      "20:\tlearn: 0.6869278\ttotal: 128ms\tremaining: 481ms\n",
      "21:\tlearn: 0.6868534\ttotal: 131ms\tremaining: 464ms\n",
      "22:\tlearn: 0.6866896\ttotal: 134ms\tremaining: 449ms\n",
      "23:\tlearn: 0.6866272\ttotal: 137ms\tremaining: 435ms\n",
      "24:\tlearn: 0.6866167\ttotal: 140ms\tremaining: 421ms\n",
      "25:\tlearn: 0.6865126\ttotal: 143ms\tremaining: 408ms\n",
      "26:\tlearn: 0.6864928\ttotal: 146ms\tremaining: 396ms\n",
      "27:\tlearn: 0.6864668\ttotal: 150ms\tremaining: 385ms\n",
      "28:\tlearn: 0.6864283\ttotal: 153ms\tremaining: 374ms\n",
      "29:\tlearn: 0.6863546\ttotal: 156ms\tremaining: 363ms\n",
      "30:\tlearn: 0.6862390\ttotal: 159ms\tremaining: 353ms\n",
      "31:\tlearn: 0.6861684\ttotal: 162ms\tremaining: 344ms\n",
      "32:\tlearn: 0.6860882\ttotal: 165ms\tremaining: 335ms\n",
      "33:\tlearn: 0.6860072\ttotal: 168ms\tremaining: 327ms\n",
      "34:\tlearn: 0.6856798\ttotal: 172ms\tremaining: 319ms\n",
      "35:\tlearn: 0.6856070\ttotal: 179ms\tremaining: 318ms\n",
      "36:\tlearn: 0.6854991\ttotal: 184ms\tremaining: 314ms\n",
      "37:\tlearn: 0.6854478\ttotal: 188ms\tremaining: 307ms\n",
      "38:\tlearn: 0.6853829\ttotal: 191ms\tremaining: 299ms\n",
      "39:\tlearn: 0.6852839\ttotal: 195ms\tremaining: 292ms\n",
      "40:\tlearn: 0.6851935\ttotal: 198ms\tremaining: 285ms\n",
      "41:\tlearn: 0.6851467\ttotal: 201ms\tremaining: 278ms\n",
      "42:\tlearn: 0.6850474\ttotal: 204ms\tremaining: 271ms\n",
      "43:\tlearn: 0.6849660\ttotal: 207ms\tremaining: 264ms\n",
      "44:\tlearn: 0.6848860\ttotal: 211ms\tremaining: 258ms\n",
      "45:\tlearn: 0.6847069\ttotal: 214ms\tremaining: 251ms\n",
      "46:\tlearn: 0.6846205\ttotal: 217ms\tremaining: 244ms\n",
      "47:\tlearn: 0.6845274\ttotal: 220ms\tremaining: 238ms\n",
      "48:\tlearn: 0.6844946\ttotal: 223ms\tremaining: 232ms\n",
      "49:\tlearn: 0.6844525\ttotal: 226ms\tremaining: 226ms\n",
      "50:\tlearn: 0.6843895\ttotal: 230ms\tremaining: 221ms\n",
      "51:\tlearn: 0.6843150\ttotal: 233ms\tremaining: 215ms\n",
      "52:\tlearn: 0.6841960\ttotal: 236ms\tremaining: 209ms\n",
      "53:\tlearn: 0.6841272\ttotal: 239ms\tremaining: 204ms\n",
      "54:\tlearn: 0.6840828\ttotal: 243ms\tremaining: 198ms\n",
      "55:\tlearn: 0.6840413\ttotal: 245ms\tremaining: 193ms\n",
      "56:\tlearn: 0.6839867\ttotal: 249ms\tremaining: 188ms\n",
      "57:\tlearn: 0.6838692\ttotal: 252ms\tremaining: 182ms\n",
      "58:\tlearn: 0.6838242\ttotal: 255ms\tremaining: 177ms\n",
      "59:\tlearn: 0.6838074\ttotal: 259ms\tremaining: 173ms\n",
      "60:\tlearn: 0.6836984\ttotal: 266ms\tremaining: 170ms\n",
      "61:\tlearn: 0.6836711\ttotal: 271ms\tremaining: 166ms\n",
      "62:\tlearn: 0.6836585\ttotal: 274ms\tremaining: 161ms\n",
      "63:\tlearn: 0.6835945\ttotal: 277ms\tremaining: 156ms\n",
      "64:\tlearn: 0.6835858\ttotal: 280ms\tremaining: 151ms\n",
      "65:\tlearn: 0.6835565\ttotal: 283ms\tremaining: 146ms\n",
      "66:\tlearn: 0.6834844\ttotal: 286ms\tremaining: 141ms\n",
      "67:\tlearn: 0.6833893\ttotal: 290ms\tremaining: 136ms\n",
      "68:\tlearn: 0.6832967\ttotal: 293ms\tremaining: 132ms\n",
      "69:\tlearn: 0.6832613\ttotal: 296ms\tremaining: 127ms\n",
      "70:\tlearn: 0.6832495\ttotal: 299ms\tremaining: 122ms\n",
      "71:\tlearn: 0.6831934\ttotal: 302ms\tremaining: 118ms\n",
      "72:\tlearn: 0.6831631\ttotal: 305ms\tremaining: 113ms\n",
      "73:\tlearn: 0.6830763\ttotal: 309ms\tremaining: 108ms\n",
      "74:\tlearn: 0.6829981\ttotal: 312ms\tremaining: 104ms\n",
      "75:\tlearn: 0.6829461\ttotal: 315ms\tremaining: 99.5ms\n",
      "76:\tlearn: 0.6828901\ttotal: 318ms\tremaining: 95ms\n",
      "77:\tlearn: 0.6828187\ttotal: 322ms\tremaining: 90.9ms\n",
      "78:\tlearn: 0.6827702\ttotal: 325ms\tremaining: 86.5ms\n",
      "79:\tlearn: 0.6826842\ttotal: 329ms\tremaining: 82.1ms\n",
      "80:\tlearn: 0.6826092\ttotal: 332ms\tremaining: 77.8ms\n",
      "81:\tlearn: 0.6825664\ttotal: 335ms\tremaining: 73.6ms\n",
      "82:\tlearn: 0.6824741\ttotal: 338ms\tremaining: 69.3ms\n",
      "83:\tlearn: 0.6824430\ttotal: 341ms\tremaining: 65ms\n",
      "84:\tlearn: 0.6824049\ttotal: 351ms\tremaining: 61.9ms\n",
      "85:\tlearn: 0.6823467\ttotal: 355ms\tremaining: 57.8ms\n",
      "86:\tlearn: 0.6822681\ttotal: 359ms\tremaining: 53.6ms\n",
      "87:\tlearn: 0.6822215\ttotal: 362ms\tremaining: 49.3ms\n",
      "88:\tlearn: 0.6822022\ttotal: 365ms\tremaining: 45.1ms\n",
      "89:\tlearn: 0.6821613\ttotal: 368ms\tremaining: 40.9ms\n",
      "90:\tlearn: 0.6821412\ttotal: 371ms\tremaining: 36.7ms\n",
      "91:\tlearn: 0.6821136\ttotal: 375ms\tremaining: 32.6ms\n",
      "92:\tlearn: 0.6820603\ttotal: 378ms\tremaining: 28.4ms\n",
      "93:\tlearn: 0.6820100\ttotal: 381ms\tremaining: 24.3ms\n",
      "94:\tlearn: 0.6819700\ttotal: 384ms\tremaining: 20.2ms\n",
      "95:\tlearn: 0.6819077\ttotal: 387ms\tremaining: 16.1ms\n",
      "96:\tlearn: 0.6818720\ttotal: 390ms\tremaining: 12.1ms\n",
      "97:\tlearn: 0.6818148\ttotal: 394ms\tremaining: 8.04ms\n",
      "98:\tlearn: 0.6817653\ttotal: 398ms\tremaining: 4.02ms\n",
      "99:\tlearn: 0.6817247\ttotal: 401ms\tremaining: 0us\n",
      "Learning rate set to 0.402147\n",
      "0:\tlearn: 0.6907917\ttotal: 3.89ms\tremaining: 385ms\n",
      "1:\tlearn: 0.6895893\ttotal: 7.36ms\tremaining: 361ms\n",
      "2:\tlearn: 0.6890026\ttotal: 10.5ms\tremaining: 340ms\n",
      "3:\tlearn: 0.6886648\ttotal: 13.5ms\tremaining: 324ms\n",
      "4:\tlearn: 0.6883906\ttotal: 16.5ms\tremaining: 313ms\n",
      "5:\tlearn: 0.6883481\ttotal: 19ms\tremaining: 298ms\n",
      "6:\tlearn: 0.6882057\ttotal: 22.1ms\tremaining: 293ms\n",
      "7:\tlearn: 0.6881161\ttotal: 25.3ms\tremaining: 291ms\n",
      "8:\tlearn: 0.6880434\ttotal: 28.3ms\tremaining: 286ms\n",
      "9:\tlearn: 0.6879450\ttotal: 31.3ms\tremaining: 281ms\n",
      "10:\tlearn: 0.6879151\ttotal: 34.3ms\tremaining: 278ms\n",
      "11:\tlearn: 0.6878572\ttotal: 37.5ms\tremaining: 275ms\n",
      "12:\tlearn: 0.6878006\ttotal: 40.5ms\tremaining: 271ms\n",
      "13:\tlearn: 0.6877180\ttotal: 43.5ms\tremaining: 267ms\n",
      "14:\tlearn: 0.6876697\ttotal: 48ms\tremaining: 272ms\n",
      "15:\tlearn: 0.6875840\ttotal: 54.3ms\tremaining: 285ms\n",
      "16:\tlearn: 0.6875375\ttotal: 57.6ms\tremaining: 281ms\n",
      "17:\tlearn: 0.6875153\ttotal: 60.8ms\tremaining: 277ms\n",
      "18:\tlearn: 0.6873818\ttotal: 64ms\tremaining: 273ms\n",
      "19:\tlearn: 0.6873092\ttotal: 67.2ms\tremaining: 269ms\n",
      "20:\tlearn: 0.6872653\ttotal: 70.3ms\tremaining: 264ms\n",
      "21:\tlearn: 0.6872091\ttotal: 73.3ms\tremaining: 260ms\n",
      "22:\tlearn: 0.6871366\ttotal: 76.4ms\tremaining: 256ms\n",
      "23:\tlearn: 0.6870961\ttotal: 79.4ms\tremaining: 251ms\n",
      "24:\tlearn: 0.6870443\ttotal: 82.3ms\tremaining: 247ms\n",
      "25:\tlearn: 0.6870197\ttotal: 85.1ms\tremaining: 242ms\n",
      "26:\tlearn: 0.6869958\ttotal: 88.4ms\tremaining: 239ms\n",
      "27:\tlearn: 0.6869164\ttotal: 91.5ms\tremaining: 235ms\n",
      "28:\tlearn: 0.6868908\ttotal: 94.5ms\tremaining: 231ms\n",
      "29:\tlearn: 0.6867413\ttotal: 97.8ms\tremaining: 228ms\n",
      "30:\tlearn: 0.6865805\ttotal: 101ms\tremaining: 225ms\n",
      "31:\tlearn: 0.6864704\ttotal: 104ms\tremaining: 221ms\n",
      "32:\tlearn: 0.6862952\ttotal: 107ms\tremaining: 218ms\n",
      "33:\tlearn: 0.6862536\ttotal: 111ms\tremaining: 215ms\n",
      "34:\tlearn: 0.6861706\ttotal: 114ms\tremaining: 211ms\n",
      "35:\tlearn: 0.6860039\ttotal: 117ms\tremaining: 208ms\n",
      "36:\tlearn: 0.6857742\ttotal: 120ms\tremaining: 205ms\n",
      "37:\tlearn: 0.6856582\ttotal: 124ms\tremaining: 202ms\n",
      "38:\tlearn: 0.6855893\ttotal: 127ms\tremaining: 198ms\n",
      "39:\tlearn: 0.6855137\ttotal: 131ms\tremaining: 196ms\n",
      "40:\tlearn: 0.6853732\ttotal: 135ms\tremaining: 195ms\n",
      "41:\tlearn: 0.6852694\ttotal: 139ms\tremaining: 192ms\n",
      "42:\tlearn: 0.6852171\ttotal: 142ms\tremaining: 188ms\n",
      "43:\tlearn: 0.6851230\ttotal: 145ms\tremaining: 185ms\n",
      "44:\tlearn: 0.6850292\ttotal: 148ms\tremaining: 181ms\n",
      "45:\tlearn: 0.6849054\ttotal: 152ms\tremaining: 178ms\n",
      "46:\tlearn: 0.6848113\ttotal: 155ms\tremaining: 175ms\n",
      "47:\tlearn: 0.6847230\ttotal: 158ms\tremaining: 171ms\n",
      "48:\tlearn: 0.6846356\ttotal: 161ms\tremaining: 168ms\n",
      "49:\tlearn: 0.6845801\ttotal: 164ms\tremaining: 164ms\n",
      "50:\tlearn: 0.6844797\ttotal: 168ms\tremaining: 161ms\n",
      "51:\tlearn: 0.6844459\ttotal: 171ms\tremaining: 158ms\n",
      "52:\tlearn: 0.6843668\ttotal: 174ms\tremaining: 155ms\n",
      "53:\tlearn: 0.6842870\ttotal: 178ms\tremaining: 151ms\n",
      "54:\tlearn: 0.6842088\ttotal: 181ms\tremaining: 148ms\n",
      "55:\tlearn: 0.6841536\ttotal: 184ms\tremaining: 145ms\n",
      "56:\tlearn: 0.6840907\ttotal: 187ms\tremaining: 141ms\n",
      "57:\tlearn: 0.6840205\ttotal: 190ms\tremaining: 138ms\n",
      "58:\tlearn: 0.6839022\ttotal: 194ms\tremaining: 135ms\n",
      "59:\tlearn: 0.6838687\ttotal: 197ms\tremaining: 131ms\n",
      "60:\tlearn: 0.6838348\ttotal: 200ms\tremaining: 128ms\n",
      "61:\tlearn: 0.6838022\ttotal: 203ms\tremaining: 124ms\n",
      "62:\tlearn: 0.6837062\ttotal: 206ms\tremaining: 121ms\n",
      "63:\tlearn: 0.6836362\ttotal: 209ms\tremaining: 118ms\n",
      "64:\tlearn: 0.6835577\ttotal: 214ms\tremaining: 115ms\n",
      "65:\tlearn: 0.6835092\ttotal: 222ms\tremaining: 115ms\n",
      "66:\tlearn: 0.6834335\ttotal: 226ms\tremaining: 111ms\n",
      "67:\tlearn: 0.6833438\ttotal: 229ms\tremaining: 108ms\n",
      "68:\tlearn: 0.6833039\ttotal: 232ms\tremaining: 104ms\n",
      "69:\tlearn: 0.6832399\ttotal: 235ms\tremaining: 101ms\n",
      "70:\tlearn: 0.6832030\ttotal: 238ms\tremaining: 97.3ms\n",
      "71:\tlearn: 0.6831591\ttotal: 241ms\tremaining: 93.8ms\n",
      "72:\tlearn: 0.6831119\ttotal: 245ms\tremaining: 90.5ms\n",
      "73:\tlearn: 0.6830474\ttotal: 248ms\tremaining: 87.1ms\n",
      "74:\tlearn: 0.6829751\ttotal: 252ms\tremaining: 83.9ms\n",
      "75:\tlearn: 0.6829007\ttotal: 256ms\tremaining: 80.7ms\n",
      "76:\tlearn: 0.6828085\ttotal: 259ms\tremaining: 77.2ms\n",
      "77:\tlearn: 0.6826939\ttotal: 262ms\tremaining: 73.8ms\n",
      "78:\tlearn: 0.6826517\ttotal: 265ms\tremaining: 70.6ms\n",
      "79:\tlearn: 0.6825662\ttotal: 270ms\tremaining: 67.4ms\n",
      "80:\tlearn: 0.6825143\ttotal: 273ms\tremaining: 64ms\n",
      "81:\tlearn: 0.6825025\ttotal: 276ms\tremaining: 60.5ms\n",
      "82:\tlearn: 0.6824550\ttotal: 279ms\tremaining: 57.1ms\n",
      "83:\tlearn: 0.6824384\ttotal: 282ms\tremaining: 53.6ms\n",
      "84:\tlearn: 0.6823434\ttotal: 285ms\tremaining: 50.3ms\n",
      "85:\tlearn: 0.6822789\ttotal: 288ms\tremaining: 46.9ms\n",
      "86:\tlearn: 0.6822635\ttotal: 291ms\tremaining: 43.5ms\n",
      "87:\tlearn: 0.6822102\ttotal: 294ms\tremaining: 40.1ms\n",
      "88:\tlearn: 0.6821043\ttotal: 297ms\tremaining: 36.8ms\n",
      "89:\tlearn: 0.6820725\ttotal: 303ms\tremaining: 33.7ms\n",
      "90:\tlearn: 0.6820333\ttotal: 310ms\tremaining: 30.6ms\n",
      "91:\tlearn: 0.6820254\ttotal: 313ms\tremaining: 27.2ms\n",
      "92:\tlearn: 0.6819799\ttotal: 316ms\tremaining: 23.8ms\n",
      "93:\tlearn: 0.6818864\ttotal: 319ms\tremaining: 20.4ms\n",
      "94:\tlearn: 0.6818245\ttotal: 322ms\tremaining: 16.9ms\n",
      "95:\tlearn: 0.6817621\ttotal: 326ms\tremaining: 13.6ms\n",
      "96:\tlearn: 0.6817639\ttotal: 328ms\tremaining: 10.2ms\n",
      "97:\tlearn: 0.6817427\ttotal: 332ms\tremaining: 6.77ms\n",
      "98:\tlearn: 0.6817118\ttotal: 335ms\tremaining: 3.38ms\n",
      "99:\tlearn: 0.6816597\ttotal: 337ms\tremaining: 0us\n",
      "Learning rate set to 0.212981\n",
      "0:\tlearn: 0.6915286\ttotal: 5.9ms\tremaining: 1.17s\n",
      "1:\tlearn: 0.6905268\ttotal: 12.3ms\tremaining: 1.22s\n",
      "2:\tlearn: 0.6897735\ttotal: 21.8ms\tremaining: 1.43s\n",
      "3:\tlearn: 0.6891313\ttotal: 31.9ms\tremaining: 1.56s\n",
      "4:\tlearn: 0.6886921\ttotal: 38.7ms\tremaining: 1.51s\n",
      "5:\tlearn: 0.6884368\ttotal: 45.9ms\tremaining: 1.48s\n",
      "6:\tlearn: 0.6882059\ttotal: 52.4ms\tremaining: 1.45s\n",
      "7:\tlearn: 0.6881583\ttotal: 57.6ms\tremaining: 1.38s\n",
      "8:\tlearn: 0.6880525\ttotal: 64ms\tremaining: 1.36s\n",
      "9:\tlearn: 0.6879350\ttotal: 70.6ms\tremaining: 1.34s\n",
      "10:\tlearn: 0.6878260\ttotal: 79.9ms\tremaining: 1.37s\n",
      "11:\tlearn: 0.6876431\ttotal: 93.1ms\tremaining: 1.46s\n",
      "12:\tlearn: 0.6875662\ttotal: 103ms\tremaining: 1.48s\n",
      "13:\tlearn: 0.6874603\ttotal: 111ms\tremaining: 1.47s\n",
      "14:\tlearn: 0.6873686\ttotal: 117ms\tremaining: 1.44s\n",
      "15:\tlearn: 0.6873047\ttotal: 123ms\tremaining: 1.42s\n",
      "16:\tlearn: 0.6872720\ttotal: 130ms\tremaining: 1.4s\n",
      "17:\tlearn: 0.6871418\ttotal: 137ms\tremaining: 1.38s\n",
      "18:\tlearn: 0.6870921\ttotal: 144ms\tremaining: 1.37s\n",
      "19:\tlearn: 0.6870374\ttotal: 150ms\tremaining: 1.35s\n",
      "20:\tlearn: 0.6870216\ttotal: 157ms\tremaining: 1.34s\n",
      "21:\tlearn: 0.6869442\ttotal: 164ms\tremaining: 1.33s\n",
      "22:\tlearn: 0.6869208\ttotal: 171ms\tremaining: 1.31s\n",
      "23:\tlearn: 0.6868974\ttotal: 184ms\tremaining: 1.35s\n",
      "24:\tlearn: 0.6868627\ttotal: 193ms\tremaining: 1.35s\n",
      "25:\tlearn: 0.6868197\ttotal: 200ms\tremaining: 1.34s\n",
      "26:\tlearn: 0.6867906\ttotal: 206ms\tremaining: 1.32s\n",
      "27:\tlearn: 0.6867572\ttotal: 214ms\tremaining: 1.31s\n",
      "28:\tlearn: 0.6867286\ttotal: 221ms\tremaining: 1.3s\n",
      "29:\tlearn: 0.6867073\ttotal: 228ms\tremaining: 1.29s\n",
      "30:\tlearn: 0.6865977\ttotal: 234ms\tremaining: 1.28s\n",
      "31:\tlearn: 0.6865618\ttotal: 242ms\tremaining: 1.27s\n",
      "32:\tlearn: 0.6865453\ttotal: 248ms\tremaining: 1.26s\n",
      "33:\tlearn: 0.6864963\ttotal: 254ms\tremaining: 1.24s\n",
      "34:\tlearn: 0.6864847\ttotal: 260ms\tremaining: 1.23s\n",
      "35:\tlearn: 0.6864777\ttotal: 265ms\tremaining: 1.21s\n",
      "36:\tlearn: 0.6864350\ttotal: 277ms\tremaining: 1.22s\n",
      "37:\tlearn: 0.6864094\ttotal: 284ms\tremaining: 1.21s\n",
      "38:\tlearn: 0.6863959\ttotal: 289ms\tremaining: 1.19s\n",
      "39:\tlearn: 0.6863769\ttotal: 296ms\tremaining: 1.18s\n",
      "40:\tlearn: 0.6863806\ttotal: 300ms\tremaining: 1.16s\n",
      "41:\tlearn: 0.6863433\ttotal: 307ms\tremaining: 1.15s\n",
      "42:\tlearn: 0.6863260\ttotal: 313ms\tremaining: 1.14s\n",
      "43:\tlearn: 0.6862895\ttotal: 319ms\tremaining: 1.13s\n",
      "44:\tlearn: 0.6862591\ttotal: 326ms\tremaining: 1.12s\n",
      "45:\tlearn: 0.6861603\ttotal: 332ms\tremaining: 1.11s\n",
      "46:\tlearn: 0.6861108\ttotal: 338ms\tremaining: 1.1s\n",
      "47:\tlearn: 0.6860786\ttotal: 345ms\tremaining: 1.09s\n",
      "48:\tlearn: 0.6860525\ttotal: 351ms\tremaining: 1.08s\n",
      "49:\tlearn: 0.6859663\ttotal: 357ms\tremaining: 1.07s\n",
      "50:\tlearn: 0.6859231\ttotal: 371ms\tremaining: 1.08s\n",
      "51:\tlearn: 0.6858823\ttotal: 376ms\tremaining: 1.07s\n",
      "52:\tlearn: 0.6858660\ttotal: 383ms\tremaining: 1.06s\n",
      "53:\tlearn: 0.6857557\ttotal: 389ms\tremaining: 1.05s\n",
      "54:\tlearn: 0.6856525\ttotal: 396ms\tremaining: 1.04s\n",
      "55:\tlearn: 0.6855657\ttotal: 401ms\tremaining: 1.03s\n",
      "56:\tlearn: 0.6854957\ttotal: 408ms\tremaining: 1.02s\n",
      "57:\tlearn: 0.6854519\ttotal: 414ms\tremaining: 1.01s\n",
      "58:\tlearn: 0.6853955\ttotal: 421ms\tremaining: 1s\n",
      "59:\tlearn: 0.6853316\ttotal: 427ms\tremaining: 997ms\n",
      "60:\tlearn: 0.6852391\ttotal: 434ms\tremaining: 990ms\n",
      "61:\tlearn: 0.6851693\ttotal: 441ms\tremaining: 982ms\n",
      "62:\tlearn: 0.6850483\ttotal: 447ms\tremaining: 972ms\n",
      "63:\tlearn: 0.6850190\ttotal: 455ms\tremaining: 967ms\n",
      "64:\tlearn: 0.6849613\ttotal: 463ms\tremaining: 962ms\n",
      "65:\tlearn: 0.6849113\ttotal: 470ms\tremaining: 955ms\n",
      "66:\tlearn: 0.6848113\ttotal: 477ms\tremaining: 947ms\n",
      "67:\tlearn: 0.6847012\ttotal: 484ms\tremaining: 939ms\n",
      "68:\tlearn: 0.6846530\ttotal: 490ms\tremaining: 931ms\n",
      "69:\tlearn: 0.6845638\ttotal: 497ms\tremaining: 922ms\n",
      "70:\tlearn: 0.6844958\ttotal: 503ms\tremaining: 914ms\n",
      "71:\tlearn: 0.6844736\ttotal: 508ms\tremaining: 904ms\n",
      "72:\tlearn: 0.6844235\ttotal: 516ms\tremaining: 897ms\n",
      "73:\tlearn: 0.6843587\ttotal: 524ms\tremaining: 891ms\n",
      "74:\tlearn: 0.6843318\ttotal: 530ms\tremaining: 883ms\n",
      "75:\tlearn: 0.6842774\ttotal: 536ms\tremaining: 874ms\n",
      "76:\tlearn: 0.6842529\ttotal: 544ms\tremaining: 869ms\n",
      "77:\tlearn: 0.6842249\ttotal: 553ms\tremaining: 866ms\n",
      "78:\tlearn: 0.6841538\ttotal: 562ms\tremaining: 861ms\n",
      "79:\tlearn: 0.6840776\ttotal: 568ms\tremaining: 852ms\n",
      "80:\tlearn: 0.6840191\ttotal: 573ms\tremaining: 842ms\n",
      "81:\tlearn: 0.6839522\ttotal: 579ms\tremaining: 834ms\n",
      "82:\tlearn: 0.6839199\ttotal: 586ms\tremaining: 826ms\n",
      "83:\tlearn: 0.6838678\ttotal: 591ms\tremaining: 816ms\n",
      "84:\tlearn: 0.6837878\ttotal: 597ms\tremaining: 808ms\n",
      "85:\tlearn: 0.6836883\ttotal: 603ms\tremaining: 799ms\n",
      "86:\tlearn: 0.6836093\ttotal: 609ms\tremaining: 791ms\n",
      "87:\tlearn: 0.6835654\ttotal: 615ms\tremaining: 782ms\n",
      "88:\tlearn: 0.6835048\ttotal: 622ms\tremaining: 775ms\n",
      "89:\tlearn: 0.6834846\ttotal: 628ms\tremaining: 768ms\n",
      "90:\tlearn: 0.6834258\ttotal: 642ms\tremaining: 769ms\n",
      "91:\tlearn: 0.6833670\ttotal: 648ms\tremaining: 761ms\n",
      "92:\tlearn: 0.6832877\ttotal: 655ms\tremaining: 753ms\n",
      "93:\tlearn: 0.6832251\ttotal: 661ms\tremaining: 746ms\n",
      "94:\tlearn: 0.6831766\ttotal: 667ms\tremaining: 738ms\n",
      "95:\tlearn: 0.6831500\ttotal: 674ms\tremaining: 730ms\n",
      "96:\tlearn: 0.6830993\ttotal: 679ms\tremaining: 721ms\n",
      "97:\tlearn: 0.6830508\ttotal: 685ms\tremaining: 713ms\n",
      "98:\tlearn: 0.6830069\ttotal: 691ms\tremaining: 705ms\n",
      "99:\tlearn: 0.6829576\ttotal: 698ms\tremaining: 698ms\n",
      "100:\tlearn: 0.6829330\ttotal: 704ms\tremaining: 690ms\n",
      "101:\tlearn: 0.6828905\ttotal: 711ms\tremaining: 683ms\n",
      "102:\tlearn: 0.6828422\ttotal: 717ms\tremaining: 675ms\n",
      "103:\tlearn: 0.6828122\ttotal: 723ms\tremaining: 668ms\n",
      "104:\tlearn: 0.6827719\ttotal: 734ms\tremaining: 664ms\n",
      "105:\tlearn: 0.6827217\ttotal: 741ms\tremaining: 657ms\n",
      "106:\tlearn: 0.6826768\ttotal: 747ms\tremaining: 649ms\n",
      "107:\tlearn: 0.6826172\ttotal: 753ms\tremaining: 642ms\n",
      "108:\tlearn: 0.6825481\ttotal: 760ms\tremaining: 635ms\n",
      "109:\tlearn: 0.6825082\ttotal: 767ms\tremaining: 627ms\n",
      "110:\tlearn: 0.6824507\ttotal: 773ms\tremaining: 620ms\n",
      "111:\tlearn: 0.6824044\ttotal: 780ms\tremaining: 613ms\n",
      "112:\tlearn: 0.6823740\ttotal: 786ms\tremaining: 605ms\n",
      "113:\tlearn: 0.6823411\ttotal: 792ms\tremaining: 598ms\n",
      "114:\tlearn: 0.6822624\ttotal: 798ms\tremaining: 590ms\n",
      "115:\tlearn: 0.6822328\ttotal: 804ms\tremaining: 582ms\n",
      "116:\tlearn: 0.6821952\ttotal: 810ms\tremaining: 575ms\n",
      "117:\tlearn: 0.6821607\ttotal: 817ms\tremaining: 568ms\n",
      "118:\tlearn: 0.6821315\ttotal: 828ms\tremaining: 563ms\n",
      "119:\tlearn: 0.6821047\ttotal: 836ms\tremaining: 557ms\n",
      "120:\tlearn: 0.6820759\ttotal: 844ms\tremaining: 551ms\n",
      "121:\tlearn: 0.6820282\ttotal: 851ms\tremaining: 544ms\n",
      "122:\tlearn: 0.6819959\ttotal: 857ms\tremaining: 537ms\n",
      "123:\tlearn: 0.6819459\ttotal: 864ms\tremaining: 530ms\n",
      "124:\tlearn: 0.6819307\ttotal: 871ms\tremaining: 522ms\n",
      "125:\tlearn: 0.6818593\ttotal: 877ms\tremaining: 515ms\n",
      "126:\tlearn: 0.6818370\ttotal: 884ms\tremaining: 508ms\n",
      "127:\tlearn: 0.6817208\ttotal: 891ms\tremaining: 501ms\n",
      "128:\tlearn: 0.6816671\ttotal: 897ms\tremaining: 494ms\n",
      "129:\tlearn: 0.6816371\ttotal: 903ms\tremaining: 486ms\n",
      "130:\tlearn: 0.6816055\ttotal: 912ms\tremaining: 481ms\n",
      "131:\tlearn: 0.6815756\ttotal: 922ms\tremaining: 475ms\n",
      "132:\tlearn: 0.6815331\ttotal: 928ms\tremaining: 468ms\n",
      "133:\tlearn: 0.6814985\ttotal: 935ms\tremaining: 460ms\n",
      "134:\tlearn: 0.6814832\ttotal: 940ms\tremaining: 453ms\n",
      "135:\tlearn: 0.6814423\ttotal: 947ms\tremaining: 446ms\n",
      "136:\tlearn: 0.6814256\ttotal: 953ms\tremaining: 438ms\n",
      "137:\tlearn: 0.6813524\ttotal: 960ms\tremaining: 431ms\n",
      "138:\tlearn: 0.6813022\ttotal: 967ms\tremaining: 424ms\n",
      "139:\tlearn: 0.6812608\ttotal: 973ms\tremaining: 417ms\n",
      "140:\tlearn: 0.6812241\ttotal: 980ms\tremaining: 410ms\n",
      "141:\tlearn: 0.6812101\ttotal: 985ms\tremaining: 403ms\n",
      "142:\tlearn: 0.6811767\ttotal: 992ms\tremaining: 395ms\n",
      "143:\tlearn: 0.6811621\ttotal: 998ms\tremaining: 388ms\n",
      "144:\tlearn: 0.6811188\ttotal: 1.01s\tremaining: 383ms\n",
      "145:\tlearn: 0.6810953\ttotal: 1.02s\tremaining: 376ms\n",
      "146:\tlearn: 0.6810577\ttotal: 1.02s\tremaining: 369ms\n",
      "147:\tlearn: 0.6810006\ttotal: 1.03s\tremaining: 362ms\n",
      "148:\tlearn: 0.6809712\ttotal: 1.04s\tremaining: 355ms\n",
      "149:\tlearn: 0.6809173\ttotal: 1.04s\tremaining: 347ms\n",
      "150:\tlearn: 0.6808344\ttotal: 1.05s\tremaining: 340ms\n",
      "151:\tlearn: 0.6808254\ttotal: 1.05s\tremaining: 333ms\n",
      "152:\tlearn: 0.6807851\ttotal: 1.06s\tremaining: 326ms\n",
      "153:\tlearn: 0.6807710\ttotal: 1.07s\tremaining: 319ms\n",
      "154:\tlearn: 0.6807454\ttotal: 1.07s\tremaining: 311ms\n",
      "155:\tlearn: 0.6807158\ttotal: 1.08s\tremaining: 304ms\n",
      "156:\tlearn: 0.6806734\ttotal: 1.08s\tremaining: 297ms\n",
      "157:\tlearn: 0.6806446\ttotal: 1.09s\tremaining: 290ms\n",
      "158:\tlearn: 0.6806173\ttotal: 1.1s\tremaining: 285ms\n",
      "159:\tlearn: 0.6805644\ttotal: 1.11s\tremaining: 278ms\n",
      "160:\tlearn: 0.6805521\ttotal: 1.12s\tremaining: 271ms\n",
      "161:\tlearn: 0.6805501\ttotal: 1.13s\tremaining: 264ms\n",
      "162:\tlearn: 0.6804823\ttotal: 1.13s\tremaining: 257ms\n",
      "163:\tlearn: 0.6804662\ttotal: 1.14s\tremaining: 250ms\n",
      "164:\tlearn: 0.6804393\ttotal: 1.14s\tremaining: 243ms\n",
      "165:\tlearn: 0.6804139\ttotal: 1.15s\tremaining: 236ms\n",
      "166:\tlearn: 0.6803931\ttotal: 1.16s\tremaining: 229ms\n",
      "167:\tlearn: 0.6803488\ttotal: 1.17s\tremaining: 222ms\n",
      "168:\tlearn: 0.6802738\ttotal: 1.17s\tremaining: 215ms\n",
      "169:\tlearn: 0.6802464\ttotal: 1.18s\tremaining: 208ms\n",
      "170:\tlearn: 0.6801855\ttotal: 1.18s\tremaining: 201ms\n",
      "171:\tlearn: 0.6801507\ttotal: 1.2s\tremaining: 195ms\n",
      "172:\tlearn: 0.6801103\ttotal: 1.21s\tremaining: 188ms\n",
      "173:\tlearn: 0.6800620\ttotal: 1.21s\tremaining: 181ms\n",
      "174:\tlearn: 0.6800352\ttotal: 1.22s\tremaining: 174ms\n",
      "175:\tlearn: 0.6800104\ttotal: 1.22s\tremaining: 167ms\n",
      "176:\tlearn: 0.6800145\ttotal: 1.23s\tremaining: 160ms\n",
      "177:\tlearn: 0.6799955\ttotal: 1.24s\tremaining: 153ms\n",
      "178:\tlearn: 0.6799565\ttotal: 1.24s\tremaining: 146ms\n",
      "179:\tlearn: 0.6799345\ttotal: 1.25s\tremaining: 139ms\n",
      "180:\tlearn: 0.6799076\ttotal: 1.26s\tremaining: 132ms\n",
      "181:\tlearn: 0.6798584\ttotal: 1.26s\tremaining: 125ms\n",
      "182:\tlearn: 0.6798531\ttotal: 1.27s\tremaining: 118ms\n",
      "183:\tlearn: 0.6798109\ttotal: 1.27s\tremaining: 111ms\n",
      "184:\tlearn: 0.6797771\ttotal: 1.28s\tremaining: 104ms\n",
      "185:\tlearn: 0.6797454\ttotal: 1.29s\tremaining: 97.3ms\n",
      "186:\tlearn: 0.6797202\ttotal: 1.3s\tremaining: 90.3ms\n",
      "187:\tlearn: 0.6796701\ttotal: 1.3s\tremaining: 83.3ms\n",
      "188:\tlearn: 0.6796237\ttotal: 1.31s\tremaining: 76.4ms\n",
      "189:\tlearn: 0.6796027\ttotal: 1.32s\tremaining: 69.4ms\n",
      "190:\tlearn: 0.6795716\ttotal: 1.32s\tremaining: 62.4ms\n",
      "191:\tlearn: 0.6795294\ttotal: 1.33s\tremaining: 55.5ms\n",
      "192:\tlearn: 0.6795048\ttotal: 1.34s\tremaining: 48.5ms\n",
      "193:\tlearn: 0.6794826\ttotal: 1.34s\tremaining: 41.6ms\n",
      "194:\tlearn: 0.6794573\ttotal: 1.35s\tremaining: 34.6ms\n",
      "195:\tlearn: 0.6794346\ttotal: 1.36s\tremaining: 27.7ms\n",
      "196:\tlearn: 0.6794106\ttotal: 1.36s\tremaining: 20.8ms\n",
      "197:\tlearn: 0.6794042\ttotal: 1.37s\tremaining: 13.8ms\n",
      "198:\tlearn: 0.6793806\ttotal: 1.38s\tremaining: 6.94ms\n",
      "199:\tlearn: 0.6793456\ttotal: 1.39s\tremaining: 0us\n",
      "Learning rate set to 0.212981\n",
      "0:\tlearn: 0.6916807\ttotal: 6.48ms\tremaining: 1.29s\n",
      "1:\tlearn: 0.6907554\ttotal: 13.3ms\tremaining: 1.31s\n",
      "2:\tlearn: 0.6902407\ttotal: 19.7ms\tremaining: 1.29s\n",
      "3:\tlearn: 0.6897056\ttotal: 25ms\tremaining: 1.23s\n",
      "4:\tlearn: 0.6893490\ttotal: 31.3ms\tremaining: 1.22s\n",
      "5:\tlearn: 0.6891035\ttotal: 37.6ms\tremaining: 1.22s\n",
      "6:\tlearn: 0.6889155\ttotal: 46.7ms\tremaining: 1.29s\n",
      "7:\tlearn: 0.6887418\ttotal: 55.9ms\tremaining: 1.34s\n",
      "8:\tlearn: 0.6886379\ttotal: 61.3ms\tremaining: 1.3s\n",
      "9:\tlearn: 0.6884285\ttotal: 67.7ms\tremaining: 1.29s\n",
      "10:\tlearn: 0.6883340\ttotal: 74.3ms\tremaining: 1.28s\n",
      "11:\tlearn: 0.6882222\ttotal: 81.1ms\tremaining: 1.27s\n",
      "12:\tlearn: 0.6881503\ttotal: 87.4ms\tremaining: 1.26s\n",
      "13:\tlearn: 0.6881277\ttotal: 91.8ms\tremaining: 1.22s\n",
      "14:\tlearn: 0.6880822\ttotal: 98.3ms\tremaining: 1.21s\n",
      "15:\tlearn: 0.6880310\ttotal: 105ms\tremaining: 1.2s\n",
      "16:\tlearn: 0.6879903\ttotal: 111ms\tremaining: 1.19s\n",
      "17:\tlearn: 0.6879080\ttotal: 117ms\tremaining: 1.19s\n",
      "18:\tlearn: 0.6878952\ttotal: 124ms\tremaining: 1.18s\n",
      "19:\tlearn: 0.6878831\ttotal: 128ms\tremaining: 1.16s\n",
      "20:\tlearn: 0.6878189\ttotal: 135ms\tremaining: 1.15s\n",
      "21:\tlearn: 0.6877770\ttotal: 148ms\tremaining: 1.2s\n",
      "22:\tlearn: 0.6877213\ttotal: 155ms\tremaining: 1.19s\n",
      "23:\tlearn: 0.6876773\ttotal: 161ms\tremaining: 1.18s\n",
      "24:\tlearn: 0.6876341\ttotal: 168ms\tremaining: 1.17s\n",
      "25:\tlearn: 0.6876084\ttotal: 174ms\tremaining: 1.17s\n",
      "26:\tlearn: 0.6876083\ttotal: 179ms\tremaining: 1.15s\n",
      "27:\tlearn: 0.6876018\ttotal: 185ms\tremaining: 1.14s\n",
      "28:\tlearn: 0.6875428\ttotal: 192ms\tremaining: 1.13s\n",
      "29:\tlearn: 0.6874993\ttotal: 199ms\tremaining: 1.13s\n",
      "30:\tlearn: 0.6874427\ttotal: 205ms\tremaining: 1.12s\n",
      "31:\tlearn: 0.6873790\ttotal: 211ms\tremaining: 1.11s\n",
      "32:\tlearn: 0.6873259\ttotal: 217ms\tremaining: 1.1s\n",
      "33:\tlearn: 0.6872555\ttotal: 224ms\tremaining: 1.09s\n",
      "34:\tlearn: 0.6871919\ttotal: 232ms\tremaining: 1.09s\n",
      "35:\tlearn: 0.6871694\ttotal: 242ms\tremaining: 1.1s\n",
      "36:\tlearn: 0.6870994\ttotal: 248ms\tremaining: 1.09s\n",
      "37:\tlearn: 0.6870393\ttotal: 254ms\tremaining: 1.08s\n",
      "38:\tlearn: 0.6869737\ttotal: 260ms\tremaining: 1.07s\n",
      "39:\tlearn: 0.6869612\ttotal: 266ms\tremaining: 1.06s\n",
      "40:\tlearn: 0.6869349\ttotal: 272ms\tremaining: 1.05s\n",
      "41:\tlearn: 0.6869124\ttotal: 278ms\tremaining: 1.04s\n",
      "42:\tlearn: 0.6868790\ttotal: 283ms\tremaining: 1.03s\n",
      "43:\tlearn: 0.6868667\ttotal: 289ms\tremaining: 1.02s\n",
      "44:\tlearn: 0.6868188\ttotal: 296ms\tremaining: 1.02s\n",
      "45:\tlearn: 0.6868234\ttotal: 301ms\tremaining: 1.01s\n",
      "46:\tlearn: 0.6867633\ttotal: 307ms\tremaining: 998ms\n",
      "47:\tlearn: 0.6867070\ttotal: 313ms\tremaining: 991ms\n",
      "48:\tlearn: 0.6866545\ttotal: 319ms\tremaining: 984ms\n",
      "49:\tlearn: 0.6865463\ttotal: 331ms\tremaining: 992ms\n",
      "50:\tlearn: 0.6864684\ttotal: 341ms\tremaining: 996ms\n",
      "51:\tlearn: 0.6864378\ttotal: 347ms\tremaining: 987ms\n",
      "52:\tlearn: 0.6863659\ttotal: 353ms\tremaining: 978ms\n",
      "53:\tlearn: 0.6863267\ttotal: 359ms\tremaining: 971ms\n",
      "54:\tlearn: 0.6862606\ttotal: 365ms\tremaining: 962ms\n",
      "55:\tlearn: 0.6861687\ttotal: 371ms\tremaining: 955ms\n",
      "56:\tlearn: 0.6861210\ttotal: 378ms\tremaining: 948ms\n",
      "57:\tlearn: 0.6860521\ttotal: 383ms\tremaining: 939ms\n",
      "58:\tlearn: 0.6860346\ttotal: 390ms\tremaining: 932ms\n",
      "59:\tlearn: 0.6859411\ttotal: 396ms\tremaining: 924ms\n",
      "60:\tlearn: 0.6858277\ttotal: 401ms\tremaining: 914ms\n",
      "61:\tlearn: 0.6857168\ttotal: 406ms\tremaining: 905ms\n",
      "62:\tlearn: 0.6856458\ttotal: 413ms\tremaining: 898ms\n",
      "63:\tlearn: 0.6856069\ttotal: 425ms\tremaining: 903ms\n",
      "64:\tlearn: 0.6854998\ttotal: 433ms\tremaining: 900ms\n",
      "65:\tlearn: 0.6854387\ttotal: 439ms\tremaining: 891ms\n",
      "66:\tlearn: 0.6853864\ttotal: 445ms\tremaining: 883ms\n",
      "67:\tlearn: 0.6853212\ttotal: 452ms\tremaining: 877ms\n",
      "68:\tlearn: 0.6852293\ttotal: 458ms\tremaining: 869ms\n",
      "69:\tlearn: 0.6852054\ttotal: 464ms\tremaining: 862ms\n",
      "70:\tlearn: 0.6851396\ttotal: 470ms\tremaining: 854ms\n",
      "71:\tlearn: 0.6850902\ttotal: 478ms\tremaining: 850ms\n",
      "72:\tlearn: 0.6850407\ttotal: 486ms\tremaining: 846ms\n",
      "73:\tlearn: 0.6849844\ttotal: 492ms\tremaining: 837ms\n",
      "74:\tlearn: 0.6849434\ttotal: 498ms\tremaining: 831ms\n",
      "75:\tlearn: 0.6848535\ttotal: 505ms\tremaining: 824ms\n",
      "76:\tlearn: 0.6847756\ttotal: 515ms\tremaining: 823ms\n",
      "77:\tlearn: 0.6847028\ttotal: 525ms\tremaining: 820ms\n",
      "78:\tlearn: 0.6846210\ttotal: 531ms\tremaining: 814ms\n",
      "79:\tlearn: 0.6845476\ttotal: 538ms\tremaining: 807ms\n",
      "80:\tlearn: 0.6845083\ttotal: 544ms\tremaining: 799ms\n",
      "81:\tlearn: 0.6844413\ttotal: 549ms\tremaining: 790ms\n",
      "82:\tlearn: 0.6844145\ttotal: 555ms\tremaining: 782ms\n",
      "83:\tlearn: 0.6843286\ttotal: 561ms\tremaining: 775ms\n",
      "84:\tlearn: 0.6842688\ttotal: 567ms\tremaining: 768ms\n",
      "85:\tlearn: 0.6842327\ttotal: 572ms\tremaining: 759ms\n",
      "86:\tlearn: 0.6841612\ttotal: 579ms\tremaining: 752ms\n",
      "87:\tlearn: 0.6840943\ttotal: 586ms\tremaining: 745ms\n",
      "88:\tlearn: 0.6840490\ttotal: 591ms\tremaining: 737ms\n",
      "89:\tlearn: 0.6839981\ttotal: 598ms\tremaining: 730ms\n",
      "90:\tlearn: 0.6839490\ttotal: 606ms\tremaining: 726ms\n",
      "91:\tlearn: 0.6839322\ttotal: 616ms\tremaining: 723ms\n",
      "92:\tlearn: 0.6838584\ttotal: 622ms\tremaining: 715ms\n",
      "93:\tlearn: 0.6838301\ttotal: 627ms\tremaining: 707ms\n",
      "94:\tlearn: 0.6837900\ttotal: 633ms\tremaining: 700ms\n",
      "95:\tlearn: 0.6837569\ttotal: 639ms\tremaining: 693ms\n",
      "96:\tlearn: 0.6836957\ttotal: 646ms\tremaining: 686ms\n",
      "97:\tlearn: 0.6836218\ttotal: 652ms\tremaining: 679ms\n",
      "98:\tlearn: 0.6835482\ttotal: 658ms\tremaining: 672ms\n",
      "99:\tlearn: 0.6834755\ttotal: 665ms\tremaining: 665ms\n",
      "100:\tlearn: 0.6834027\ttotal: 671ms\tremaining: 658ms\n",
      "101:\tlearn: 0.6833334\ttotal: 677ms\tremaining: 650ms\n",
      "102:\tlearn: 0.6832741\ttotal: 683ms\tremaining: 643ms\n",
      "103:\tlearn: 0.6832001\ttotal: 689ms\tremaining: 636ms\n",
      "104:\tlearn: 0.6831668\ttotal: 698ms\tremaining: 632ms\n",
      "105:\tlearn: 0.6831224\ttotal: 709ms\tremaining: 629ms\n",
      "106:\tlearn: 0.6830343\ttotal: 716ms\tremaining: 622ms\n",
      "107:\tlearn: 0.6829434\ttotal: 722ms\tremaining: 615ms\n",
      "108:\tlearn: 0.6828877\ttotal: 728ms\tremaining: 608ms\n",
      "109:\tlearn: 0.6828582\ttotal: 734ms\tremaining: 600ms\n",
      "110:\tlearn: 0.6827910\ttotal: 740ms\tremaining: 593ms\n",
      "111:\tlearn: 0.6827518\ttotal: 746ms\tremaining: 586ms\n",
      "112:\tlearn: 0.6827057\ttotal: 753ms\tremaining: 580ms\n",
      "113:\tlearn: 0.6826467\ttotal: 760ms\tremaining: 573ms\n",
      "114:\tlearn: 0.6826018\ttotal: 767ms\tremaining: 567ms\n",
      "115:\tlearn: 0.6825129\ttotal: 773ms\tremaining: 560ms\n",
      "116:\tlearn: 0.6824320\ttotal: 779ms\tremaining: 552ms\n",
      "117:\tlearn: 0.6823915\ttotal: 784ms\tremaining: 545ms\n",
      "118:\tlearn: 0.6823420\ttotal: 792ms\tremaining: 539ms\n",
      "119:\tlearn: 0.6823034\ttotal: 802ms\tremaining: 534ms\n",
      "120:\tlearn: 0.6822541\ttotal: 808ms\tremaining: 528ms\n",
      "121:\tlearn: 0.6822235\ttotal: 815ms\tremaining: 521ms\n",
      "122:\tlearn: 0.6821869\ttotal: 820ms\tremaining: 514ms\n",
      "123:\tlearn: 0.6821402\ttotal: 826ms\tremaining: 506ms\n",
      "124:\tlearn: 0.6821186\ttotal: 833ms\tremaining: 500ms\n",
      "125:\tlearn: 0.6820610\ttotal: 839ms\tremaining: 493ms\n",
      "126:\tlearn: 0.6820463\ttotal: 845ms\tremaining: 486ms\n",
      "127:\tlearn: 0.6820179\ttotal: 852ms\tremaining: 479ms\n",
      "128:\tlearn: 0.6820108\ttotal: 859ms\tremaining: 473ms\n",
      "129:\tlearn: 0.6819765\ttotal: 866ms\tremaining: 466ms\n",
      "130:\tlearn: 0.6819350\ttotal: 872ms\tremaining: 459ms\n",
      "131:\tlearn: 0.6818863\ttotal: 878ms\tremaining: 452ms\n",
      "132:\tlearn: 0.6818412\ttotal: 888ms\tremaining: 448ms\n",
      "133:\tlearn: 0.6818304\ttotal: 897ms\tremaining: 442ms\n",
      "134:\tlearn: 0.6817840\ttotal: 904ms\tremaining: 435ms\n",
      "135:\tlearn: 0.6817486\ttotal: 911ms\tremaining: 429ms\n",
      "136:\tlearn: 0.6816819\ttotal: 917ms\tremaining: 422ms\n",
      "137:\tlearn: 0.6816629\ttotal: 923ms\tremaining: 415ms\n",
      "138:\tlearn: 0.6816066\ttotal: 930ms\tremaining: 408ms\n",
      "139:\tlearn: 0.6815713\ttotal: 936ms\tremaining: 401ms\n",
      "140:\tlearn: 0.6814507\ttotal: 942ms\tremaining: 394ms\n",
      "141:\tlearn: 0.6813889\ttotal: 949ms\tremaining: 388ms\n",
      "142:\tlearn: 0.6813469\ttotal: 955ms\tremaining: 381ms\n",
      "143:\tlearn: 0.6813331\ttotal: 961ms\tremaining: 374ms\n",
      "144:\tlearn: 0.6813095\ttotal: 966ms\tremaining: 367ms\n",
      "145:\tlearn: 0.6812852\ttotal: 973ms\tremaining: 360ms\n",
      "146:\tlearn: 0.6812538\ttotal: 985ms\tremaining: 355ms\n",
      "147:\tlearn: 0.6811916\ttotal: 993ms\tremaining: 349ms\n",
      "148:\tlearn: 0.6811731\ttotal: 999ms\tremaining: 342ms\n",
      "149:\tlearn: 0.6811614\ttotal: 1s\tremaining: 335ms\n",
      "150:\tlearn: 0.6811280\ttotal: 1.01s\tremaining: 328ms\n",
      "151:\tlearn: 0.6810792\ttotal: 1.02s\tremaining: 321ms\n",
      "152:\tlearn: 0.6810590\ttotal: 1.02s\tremaining: 314ms\n",
      "153:\tlearn: 0.6810180\ttotal: 1.03s\tremaining: 307ms\n",
      "154:\tlearn: 0.6809946\ttotal: 1.03s\tremaining: 300ms\n",
      "155:\tlearn: 0.6809619\ttotal: 1.04s\tremaining: 293ms\n",
      "156:\tlearn: 0.6809114\ttotal: 1.05s\tremaining: 287ms\n",
      "157:\tlearn: 0.6808884\ttotal: 1.05s\tremaining: 280ms\n",
      "158:\tlearn: 0.6808527\ttotal: 1.06s\tremaining: 273ms\n",
      "159:\tlearn: 0.6807958\ttotal: 1.06s\tremaining: 266ms\n",
      "160:\tlearn: 0.6807861\ttotal: 1.08s\tremaining: 261ms\n",
      "161:\tlearn: 0.6807554\ttotal: 1.08s\tremaining: 255ms\n",
      "162:\tlearn: 0.6807074\ttotal: 1.09s\tremaining: 248ms\n",
      "163:\tlearn: 0.6806649\ttotal: 1.1s\tremaining: 241ms\n",
      "164:\tlearn: 0.6806603\ttotal: 1.1s\tremaining: 234ms\n",
      "165:\tlearn: 0.6806115\ttotal: 1.11s\tremaining: 227ms\n",
      "166:\tlearn: 0.6805866\ttotal: 1.11s\tremaining: 220ms\n",
      "167:\tlearn: 0.6805721\ttotal: 1.12s\tremaining: 214ms\n",
      "168:\tlearn: 0.6804959\ttotal: 1.13s\tremaining: 207ms\n",
      "169:\tlearn: 0.6804871\ttotal: 1.13s\tremaining: 200ms\n",
      "170:\tlearn: 0.6804653\ttotal: 1.14s\tremaining: 193ms\n",
      "171:\tlearn: 0.6804272\ttotal: 1.14s\tremaining: 186ms\n",
      "172:\tlearn: 0.6803840\ttotal: 1.15s\tremaining: 180ms\n",
      "173:\tlearn: 0.6803802\ttotal: 1.16s\tremaining: 173ms\n",
      "174:\tlearn: 0.6803470\ttotal: 1.17s\tremaining: 167ms\n",
      "175:\tlearn: 0.6803143\ttotal: 1.18s\tremaining: 160ms\n",
      "176:\tlearn: 0.6803049\ttotal: 1.18s\tremaining: 153ms\n",
      "177:\tlearn: 0.6802784\ttotal: 1.19s\tremaining: 147ms\n",
      "178:\tlearn: 0.6802388\ttotal: 1.19s\tremaining: 140ms\n",
      "179:\tlearn: 0.6801861\ttotal: 1.2s\tremaining: 133ms\n",
      "180:\tlearn: 0.6801466\ttotal: 1.2s\tremaining: 126ms\n",
      "181:\tlearn: 0.6801040\ttotal: 1.21s\tremaining: 120ms\n",
      "182:\tlearn: 0.6800814\ttotal: 1.22s\tremaining: 113ms\n",
      "183:\tlearn: 0.6800493\ttotal: 1.22s\tremaining: 106ms\n",
      "184:\tlearn: 0.6800239\ttotal: 1.23s\tremaining: 99.7ms\n",
      "185:\tlearn: 0.6800041\ttotal: 1.24s\tremaining: 93.1ms\n",
      "186:\tlearn: 0.6799322\ttotal: 1.24s\tremaining: 86.4ms\n",
      "187:\tlearn: 0.6798988\ttotal: 1.25s\tremaining: 79.7ms\n",
      "188:\tlearn: 0.6798848\ttotal: 1.26s\tremaining: 73.3ms\n",
      "189:\tlearn: 0.6798548\ttotal: 1.27s\tremaining: 66.8ms\n",
      "190:\tlearn: 0.6798273\ttotal: 1.27s\tremaining: 60.1ms\n",
      "191:\tlearn: 0.6798102\ttotal: 1.28s\tremaining: 53.4ms\n",
      "192:\tlearn: 0.6797833\ttotal: 1.28s\tremaining: 46.6ms\n",
      "193:\tlearn: 0.6797861\ttotal: 1.29s\tremaining: 40ms\n",
      "194:\tlearn: 0.6797066\ttotal: 1.3s\tremaining: 33.3ms\n",
      "195:\tlearn: 0.6796977\ttotal: 1.3s\tremaining: 26.6ms\n",
      "196:\tlearn: 0.6796512\ttotal: 1.31s\tremaining: 20ms\n",
      "197:\tlearn: 0.6796388\ttotal: 1.32s\tremaining: 13.3ms\n",
      "198:\tlearn: 0.6795600\ttotal: 1.32s\tremaining: 6.65ms\n",
      "199:\tlearn: 0.6795203\ttotal: 1.33s\tremaining: 0us\n",
      "Learning rate set to 0.402147\n",
      "0:\tlearn: 0.6903771\ttotal: 4.31ms\tremaining: 426ms\n",
      "1:\tlearn: 0.6887534\ttotal: 9.56ms\tremaining: 469ms\n",
      "2:\tlearn: 0.6881968\ttotal: 13.2ms\tremaining: 428ms\n",
      "3:\tlearn: 0.6876956\ttotal: 23.5ms\tremaining: 565ms\n",
      "4:\tlearn: 0.6875422\ttotal: 27.7ms\tremaining: 525ms\n",
      "5:\tlearn: 0.6872971\ttotal: 30.8ms\tremaining: 483ms\n",
      "6:\tlearn: 0.6871213\ttotal: 35.5ms\tremaining: 471ms\n",
      "7:\tlearn: 0.6869695\ttotal: 43.7ms\tremaining: 503ms\n",
      "8:\tlearn: 0.6867312\ttotal: 51.7ms\tremaining: 523ms\n",
      "9:\tlearn: 0.6864212\ttotal: 59.7ms\tremaining: 538ms\n",
      "10:\tlearn: 0.6862787\ttotal: 67.6ms\tremaining: 547ms\n",
      "11:\tlearn: 0.6860583\ttotal: 76.1ms\tremaining: 558ms\n",
      "12:\tlearn: 0.6860160\ttotal: 81ms\tremaining: 542ms\n",
      "13:\tlearn: 0.6858263\ttotal: 88.8ms\tremaining: 546ms\n",
      "14:\tlearn: 0.6857791\ttotal: 92.9ms\tremaining: 526ms\n",
      "15:\tlearn: 0.6856345\ttotal: 101ms\tremaining: 529ms\n",
      "16:\tlearn: 0.6856132\ttotal: 104ms\tremaining: 510ms\n",
      "17:\tlearn: 0.6854284\ttotal: 112ms\tremaining: 512ms\n",
      "18:\tlearn: 0.6854155\ttotal: 116ms\tremaining: 493ms\n",
      "19:\tlearn: 0.6853210\ttotal: 124ms\tremaining: 496ms\n",
      "20:\tlearn: 0.6851962\ttotal: 133ms\tremaining: 498ms\n",
      "21:\tlearn: 0.6850589\ttotal: 141ms\tremaining: 500ms\n",
      "22:\tlearn: 0.6849802\ttotal: 149ms\tremaining: 499ms\n",
      "23:\tlearn: 0.6848973\ttotal: 157ms\tremaining: 497ms\n",
      "24:\tlearn: 0.6848164\ttotal: 165ms\tremaining: 494ms\n",
      "25:\tlearn: 0.6848160\ttotal: 168ms\tremaining: 479ms\n",
      "26:\tlearn: 0.6847167\ttotal: 176ms\tremaining: 476ms\n",
      "27:\tlearn: 0.6844658\ttotal: 184ms\tremaining: 474ms\n",
      "28:\tlearn: 0.6842396\ttotal: 193ms\tremaining: 472ms\n",
      "29:\tlearn: 0.6839744\ttotal: 200ms\tremaining: 468ms\n",
      "30:\tlearn: 0.6837235\ttotal: 208ms\tremaining: 464ms\n",
      "31:\tlearn: 0.6835586\ttotal: 217ms\tremaining: 460ms\n",
      "32:\tlearn: 0.6833141\ttotal: 225ms\tremaining: 456ms\n",
      "33:\tlearn: 0.6829663\ttotal: 233ms\tremaining: 453ms\n",
      "34:\tlearn: 0.6828031\ttotal: 242ms\tremaining: 449ms\n",
      "35:\tlearn: 0.6825306\ttotal: 250ms\tremaining: 444ms\n",
      "36:\tlearn: 0.6824061\ttotal: 258ms\tremaining: 439ms\n",
      "37:\tlearn: 0.6823147\ttotal: 265ms\tremaining: 433ms\n",
      "38:\tlearn: 0.6820662\ttotal: 273ms\tremaining: 428ms\n",
      "39:\tlearn: 0.6818821\ttotal: 281ms\tremaining: 422ms\n",
      "40:\tlearn: 0.6816999\ttotal: 290ms\tremaining: 417ms\n",
      "41:\tlearn: 0.6815325\ttotal: 297ms\tremaining: 411ms\n",
      "42:\tlearn: 0.6814516\ttotal: 305ms\tremaining: 405ms\n",
      "43:\tlearn: 0.6812414\ttotal: 313ms\tremaining: 399ms\n",
      "44:\tlearn: 0.6810601\ttotal: 322ms\tremaining: 393ms\n",
      "45:\tlearn: 0.6809176\ttotal: 331ms\tremaining: 389ms\n",
      "46:\tlearn: 0.6807903\ttotal: 340ms\tremaining: 383ms\n",
      "47:\tlearn: 0.6806736\ttotal: 348ms\tremaining: 377ms\n",
      "48:\tlearn: 0.6805129\ttotal: 356ms\tremaining: 371ms\n",
      "49:\tlearn: 0.6803390\ttotal: 365ms\tremaining: 365ms\n",
      "50:\tlearn: 0.6801555\ttotal: 374ms\tremaining: 359ms\n",
      "51:\tlearn: 0.6800791\ttotal: 382ms\tremaining: 352ms\n",
      "52:\tlearn: 0.6798731\ttotal: 390ms\tremaining: 346ms\n",
      "53:\tlearn: 0.6797776\ttotal: 400ms\tremaining: 341ms\n",
      "54:\tlearn: 0.6796679\ttotal: 410ms\tremaining: 336ms\n",
      "55:\tlearn: 0.6795695\ttotal: 422ms\tremaining: 331ms\n",
      "56:\tlearn: 0.6794662\ttotal: 432ms\tremaining: 326ms\n",
      "57:\tlearn: 0.6793882\ttotal: 440ms\tremaining: 318ms\n",
      "58:\tlearn: 0.6792592\ttotal: 448ms\tremaining: 311ms\n",
      "59:\tlearn: 0.6792044\ttotal: 455ms\tremaining: 304ms\n",
      "60:\tlearn: 0.6790974\ttotal: 463ms\tremaining: 296ms\n",
      "61:\tlearn: 0.6789458\ttotal: 473ms\tremaining: 290ms\n",
      "62:\tlearn: 0.6787954\ttotal: 481ms\tremaining: 282ms\n",
      "63:\tlearn: 0.6786120\ttotal: 488ms\tremaining: 275ms\n",
      "64:\tlearn: 0.6785052\ttotal: 496ms\tremaining: 267ms\n",
      "65:\tlearn: 0.6784050\ttotal: 504ms\tremaining: 260ms\n",
      "66:\tlearn: 0.6782470\ttotal: 514ms\tremaining: 253ms\n",
      "67:\tlearn: 0.6781128\ttotal: 525ms\tremaining: 247ms\n",
      "68:\tlearn: 0.6778926\ttotal: 535ms\tremaining: 240ms\n",
      "69:\tlearn: 0.6778144\ttotal: 544ms\tremaining: 233ms\n",
      "70:\tlearn: 0.6777543\ttotal: 555ms\tremaining: 227ms\n",
      "71:\tlearn: 0.6776620\ttotal: 564ms\tremaining: 220ms\n",
      "72:\tlearn: 0.6775490\ttotal: 576ms\tremaining: 213ms\n",
      "73:\tlearn: 0.6774161\ttotal: 590ms\tremaining: 207ms\n",
      "74:\tlearn: 0.6773077\ttotal: 604ms\tremaining: 201ms\n",
      "75:\tlearn: 0.6771996\ttotal: 618ms\tremaining: 195ms\n",
      "76:\tlearn: 0.6771282\ttotal: 633ms\tremaining: 189ms\n",
      "77:\tlearn: 0.6769232\ttotal: 648ms\tremaining: 183ms\n",
      "78:\tlearn: 0.6768737\ttotal: 661ms\tremaining: 176ms\n",
      "79:\tlearn: 0.6768063\ttotal: 670ms\tremaining: 167ms\n",
      "80:\tlearn: 0.6766926\ttotal: 678ms\tremaining: 159ms\n",
      "81:\tlearn: 0.6765337\ttotal: 686ms\tremaining: 151ms\n",
      "82:\tlearn: 0.6764390\ttotal: 694ms\tremaining: 142ms\n",
      "83:\tlearn: 0.6763819\ttotal: 702ms\tremaining: 134ms\n",
      "84:\tlearn: 0.6762239\ttotal: 710ms\tremaining: 125ms\n",
      "85:\tlearn: 0.6761209\ttotal: 718ms\tremaining: 117ms\n",
      "86:\tlearn: 0.6760397\ttotal: 726ms\tremaining: 108ms\n",
      "87:\tlearn: 0.6759570\ttotal: 734ms\tremaining: 100ms\n",
      "88:\tlearn: 0.6758637\ttotal: 742ms\tremaining: 91.7ms\n",
      "89:\tlearn: 0.6757309\ttotal: 750ms\tremaining: 83.3ms\n",
      "90:\tlearn: 0.6756801\ttotal: 758ms\tremaining: 75ms\n",
      "91:\tlearn: 0.6755435\ttotal: 766ms\tremaining: 66.6ms\n",
      "92:\tlearn: 0.6754165\ttotal: 774ms\tremaining: 58.3ms\n",
      "93:\tlearn: 0.6753194\ttotal: 783ms\tremaining: 50ms\n",
      "94:\tlearn: 0.6752419\ttotal: 791ms\tremaining: 41.6ms\n",
      "95:\tlearn: 0.6751925\ttotal: 800ms\tremaining: 33.3ms\n",
      "96:\tlearn: 0.6750664\ttotal: 808ms\tremaining: 25ms\n",
      "97:\tlearn: 0.6750386\ttotal: 817ms\tremaining: 16.7ms\n",
      "98:\tlearn: 0.6749527\ttotal: 829ms\tremaining: 8.37ms\n",
      "99:\tlearn: 0.6748461\ttotal: 837ms\tremaining: 0us\n",
      "Learning rate set to 0.402147\n",
      "0:\tlearn: 0.6907917\ttotal: 4.36ms\tremaining: 432ms\n",
      "1:\tlearn: 0.6893161\ttotal: 9.25ms\tremaining: 453ms\n",
      "2:\tlearn: 0.6889336\ttotal: 12.9ms\tremaining: 417ms\n",
      "3:\tlearn: 0.6885180\ttotal: 21.5ms\tremaining: 516ms\n",
      "4:\tlearn: 0.6883537\ttotal: 25.6ms\tremaining: 486ms\n",
      "5:\tlearn: 0.6882450\ttotal: 28.6ms\tremaining: 449ms\n",
      "6:\tlearn: 0.6880814\ttotal: 31.7ms\tremaining: 421ms\n",
      "7:\tlearn: 0.6877818\ttotal: 39.6ms\tremaining: 455ms\n",
      "8:\tlearn: 0.6874377\ttotal: 47.3ms\tremaining: 478ms\n",
      "9:\tlearn: 0.6872422\ttotal: 55.8ms\tremaining: 502ms\n",
      "10:\tlearn: 0.6872427\ttotal: 58ms\tremaining: 469ms\n",
      "11:\tlearn: 0.6870571\ttotal: 65.7ms\tremaining: 482ms\n",
      "12:\tlearn: 0.6869363\ttotal: 73ms\tremaining: 489ms\n",
      "13:\tlearn: 0.6866889\ttotal: 80.8ms\tremaining: 497ms\n",
      "14:\tlearn: 0.6865546\ttotal: 88.6ms\tremaining: 502ms\n",
      "15:\tlearn: 0.6864512\ttotal: 92.6ms\tremaining: 486ms\n",
      "16:\tlearn: 0.6862283\ttotal: 101ms\tremaining: 491ms\n",
      "17:\tlearn: 0.6860613\ttotal: 108ms\tremaining: 494ms\n",
      "18:\tlearn: 0.6860177\ttotal: 113ms\tremaining: 482ms\n",
      "19:\tlearn: 0.6860160\ttotal: 115ms\tremaining: 462ms\n",
      "20:\tlearn: 0.6859730\ttotal: 119ms\tremaining: 448ms\n",
      "21:\tlearn: 0.6858365\ttotal: 127ms\tremaining: 450ms\n",
      "22:\tlearn: 0.6857535\ttotal: 135ms\tremaining: 450ms\n",
      "23:\tlearn: 0.6855998\ttotal: 142ms\tremaining: 450ms\n",
      "24:\tlearn: 0.6854502\ttotal: 150ms\tremaining: 450ms\n",
      "25:\tlearn: 0.6853052\ttotal: 158ms\tremaining: 450ms\n",
      "26:\tlearn: 0.6850519\ttotal: 166ms\tremaining: 448ms\n",
      "27:\tlearn: 0.6849086\ttotal: 173ms\tremaining: 446ms\n",
      "28:\tlearn: 0.6847720\ttotal: 182ms\tremaining: 444ms\n",
      "29:\tlearn: 0.6844628\ttotal: 190ms\tremaining: 442ms\n",
      "30:\tlearn: 0.6843038\ttotal: 198ms\tremaining: 440ms\n",
      "31:\tlearn: 0.6840497\ttotal: 206ms\tremaining: 437ms\n",
      "32:\tlearn: 0.6838066\ttotal: 213ms\tremaining: 433ms\n",
      "33:\tlearn: 0.6835534\ttotal: 221ms\tremaining: 429ms\n",
      "34:\tlearn: 0.6833062\ttotal: 229ms\tremaining: 425ms\n",
      "35:\tlearn: 0.6830375\ttotal: 236ms\tremaining: 420ms\n",
      "36:\tlearn: 0.6828919\ttotal: 244ms\tremaining: 416ms\n",
      "37:\tlearn: 0.6826224\ttotal: 252ms\tremaining: 411ms\n",
      "38:\tlearn: 0.6824839\ttotal: 260ms\tremaining: 406ms\n",
      "39:\tlearn: 0.6822109\ttotal: 267ms\tremaining: 401ms\n",
      "40:\tlearn: 0.6820741\ttotal: 275ms\tremaining: 396ms\n",
      "41:\tlearn: 0.6819374\ttotal: 283ms\tremaining: 391ms\n",
      "42:\tlearn: 0.6818635\ttotal: 291ms\tremaining: 385ms\n",
      "43:\tlearn: 0.6817315\ttotal: 298ms\tremaining: 380ms\n",
      "44:\tlearn: 0.6814982\ttotal: 306ms\tremaining: 374ms\n",
      "45:\tlearn: 0.6813860\ttotal: 314ms\tremaining: 368ms\n",
      "46:\tlearn: 0.6813066\ttotal: 322ms\tremaining: 363ms\n",
      "47:\tlearn: 0.6811111\ttotal: 329ms\tremaining: 357ms\n",
      "48:\tlearn: 0.6809939\ttotal: 337ms\tremaining: 351ms\n",
      "49:\tlearn: 0.6808048\ttotal: 345ms\tremaining: 345ms\n",
      "50:\tlearn: 0.6806434\ttotal: 352ms\tremaining: 338ms\n",
      "51:\tlearn: 0.6805383\ttotal: 360ms\tremaining: 333ms\n",
      "52:\tlearn: 0.6804128\ttotal: 368ms\tremaining: 326ms\n",
      "53:\tlearn: 0.6803303\ttotal: 376ms\tremaining: 320ms\n",
      "54:\tlearn: 0.6801831\ttotal: 384ms\tremaining: 314ms\n",
      "55:\tlearn: 0.6800609\ttotal: 393ms\tremaining: 309ms\n",
      "56:\tlearn: 0.6799766\ttotal: 401ms\tremaining: 302ms\n",
      "57:\tlearn: 0.6798102\ttotal: 408ms\tremaining: 296ms\n",
      "58:\tlearn: 0.6796600\ttotal: 416ms\tremaining: 289ms\n",
      "59:\tlearn: 0.6795939\ttotal: 424ms\tremaining: 282ms\n",
      "60:\tlearn: 0.6795087\ttotal: 432ms\tremaining: 276ms\n",
      "61:\tlearn: 0.6794027\ttotal: 440ms\tremaining: 270ms\n",
      "62:\tlearn: 0.6791974\ttotal: 447ms\tremaining: 263ms\n",
      "63:\tlearn: 0.6789736\ttotal: 455ms\tremaining: 256ms\n",
      "64:\tlearn: 0.6789055\ttotal: 463ms\tremaining: 249ms\n",
      "65:\tlearn: 0.6787742\ttotal: 471ms\tremaining: 243ms\n",
      "66:\tlearn: 0.6786720\ttotal: 479ms\tremaining: 236ms\n",
      "67:\tlearn: 0.6785600\ttotal: 488ms\tremaining: 229ms\n",
      "68:\tlearn: 0.6784215\ttotal: 495ms\tremaining: 222ms\n",
      "69:\tlearn: 0.6782564\ttotal: 503ms\tremaining: 215ms\n",
      "70:\tlearn: 0.6781912\ttotal: 510ms\tremaining: 208ms\n",
      "71:\tlearn: 0.6781246\ttotal: 518ms\tremaining: 201ms\n",
      "72:\tlearn: 0.6779501\ttotal: 526ms\tremaining: 194ms\n",
      "73:\tlearn: 0.6778190\ttotal: 533ms\tremaining: 187ms\n",
      "74:\tlearn: 0.6777088\ttotal: 541ms\tremaining: 180ms\n",
      "75:\tlearn: 0.6776317\ttotal: 549ms\tremaining: 173ms\n",
      "76:\tlearn: 0.6775770\ttotal: 556ms\tremaining: 166ms\n",
      "77:\tlearn: 0.6774557\ttotal: 564ms\tremaining: 159ms\n",
      "78:\tlearn: 0.6773266\ttotal: 572ms\tremaining: 152ms\n",
      "79:\tlearn: 0.6772003\ttotal: 579ms\tremaining: 145ms\n",
      "80:\tlearn: 0.6771172\ttotal: 592ms\tremaining: 139ms\n",
      "81:\tlearn: 0.6770710\ttotal: 600ms\tremaining: 132ms\n",
      "82:\tlearn: 0.6770129\ttotal: 608ms\tremaining: 124ms\n",
      "83:\tlearn: 0.6769338\ttotal: 615ms\tremaining: 117ms\n",
      "84:\tlearn: 0.6768693\ttotal: 623ms\tremaining: 110ms\n",
      "85:\tlearn: 0.6767532\ttotal: 631ms\tremaining: 103ms\n",
      "86:\tlearn: 0.6767186\ttotal: 638ms\tremaining: 95.4ms\n",
      "87:\tlearn: 0.6765899\ttotal: 646ms\tremaining: 88ms\n",
      "88:\tlearn: 0.6765536\ttotal: 653ms\tremaining: 80.7ms\n",
      "89:\tlearn: 0.6765186\ttotal: 661ms\tremaining: 73.5ms\n",
      "90:\tlearn: 0.6765026\ttotal: 669ms\tremaining: 66.2ms\n",
      "91:\tlearn: 0.6764425\ttotal: 678ms\tremaining: 58.9ms\n",
      "92:\tlearn: 0.6763699\ttotal: 685ms\tremaining: 51.6ms\n",
      "93:\tlearn: 0.6762644\ttotal: 693ms\tremaining: 44.2ms\n",
      "94:\tlearn: 0.6761033\ttotal: 701ms\tremaining: 36.9ms\n",
      "95:\tlearn: 0.6760314\ttotal: 708ms\tremaining: 29.5ms\n",
      "96:\tlearn: 0.6759686\ttotal: 716ms\tremaining: 22.1ms\n",
      "97:\tlearn: 0.6758734\ttotal: 724ms\tremaining: 14.8ms\n",
      "98:\tlearn: 0.6758250\ttotal: 732ms\tremaining: 7.39ms\n",
      "99:\tlearn: 0.6757087\ttotal: 742ms\tremaining: 0us\n",
      "Learning rate set to 0.212981\n",
      "0:\tlearn: 0.6915286\ttotal: 5.65ms\tremaining: 1.12s\n",
      "1:\tlearn: 0.6904903\ttotal: 10.8ms\tremaining: 1.07s\n",
      "2:\tlearn: 0.6898372\ttotal: 17.9ms\tremaining: 1.18s\n",
      "3:\tlearn: 0.6892271\ttotal: 24.9ms\tremaining: 1.22s\n",
      "4:\tlearn: 0.6885995\ttotal: 35.3ms\tremaining: 1.38s\n",
      "5:\tlearn: 0.6881521\ttotal: 48.8ms\tremaining: 1.58s\n",
      "6:\tlearn: 0.6877756\ttotal: 62.7ms\tremaining: 1.73s\n",
      "7:\tlearn: 0.6873907\ttotal: 79.9ms\tremaining: 1.92s\n",
      "8:\tlearn: 0.6870422\ttotal: 97ms\tremaining: 2.06s\n",
      "9:\tlearn: 0.6869712\ttotal: 104ms\tremaining: 1.98s\n",
      "10:\tlearn: 0.6867836\ttotal: 123ms\tremaining: 2.12s\n",
      "11:\tlearn: 0.6865681\ttotal: 140ms\tremaining: 2.19s\n",
      "12:\tlearn: 0.6864981\ttotal: 148ms\tremaining: 2.13s\n",
      "13:\tlearn: 0.6863554\ttotal: 164ms\tremaining: 2.18s\n",
      "14:\tlearn: 0.6861780\ttotal: 180ms\tremaining: 2.23s\n",
      "15:\tlearn: 0.6861136\ttotal: 189ms\tremaining: 2.17s\n",
      "16:\tlearn: 0.6860205\ttotal: 204ms\tremaining: 2.2s\n",
      "17:\tlearn: 0.6859205\ttotal: 218ms\tremaining: 2.2s\n",
      "18:\tlearn: 0.6857941\ttotal: 234ms\tremaining: 2.23s\n",
      "19:\tlearn: 0.6857814\ttotal: 245ms\tremaining: 2.21s\n",
      "20:\tlearn: 0.6857744\ttotal: 255ms\tremaining: 2.17s\n",
      "21:\tlearn: 0.6855337\ttotal: 272ms\tremaining: 2.2s\n",
      "22:\tlearn: 0.6855183\ttotal: 279ms\tremaining: 2.14s\n",
      "23:\tlearn: 0.6855135\ttotal: 283ms\tremaining: 2.07s\n",
      "24:\tlearn: 0.6853537\ttotal: 299ms\tremaining: 2.09s\n",
      "25:\tlearn: 0.6853238\ttotal: 306ms\tremaining: 2.05s\n",
      "26:\tlearn: 0.6853248\ttotal: 311ms\tremaining: 1.99s\n",
      "27:\tlearn: 0.6852068\ttotal: 325ms\tremaining: 1.99s\n",
      "28:\tlearn: 0.6850907\ttotal: 341ms\tremaining: 2.01s\n",
      "29:\tlearn: 0.6849110\ttotal: 357ms\tremaining: 2.02s\n",
      "30:\tlearn: 0.6848968\ttotal: 364ms\tremaining: 1.99s\n",
      "31:\tlearn: 0.6848755\ttotal: 372ms\tremaining: 1.95s\n",
      "32:\tlearn: 0.6847118\ttotal: 389ms\tremaining: 1.97s\n",
      "33:\tlearn: 0.6846101\ttotal: 402ms\tremaining: 1.96s\n",
      "34:\tlearn: 0.6844871\ttotal: 418ms\tremaining: 1.97s\n",
      "35:\tlearn: 0.6843774\ttotal: 434ms\tremaining: 1.98s\n",
      "36:\tlearn: 0.6843267\ttotal: 442ms\tremaining: 1.95s\n",
      "37:\tlearn: 0.6842811\ttotal: 450ms\tremaining: 1.92s\n",
      "38:\tlearn: 0.6841495\ttotal: 470ms\tremaining: 1.94s\n",
      "39:\tlearn: 0.6840972\ttotal: 477ms\tremaining: 1.91s\n",
      "40:\tlearn: 0.6839875\ttotal: 493ms\tremaining: 1.91s\n",
      "41:\tlearn: 0.6839742\ttotal: 501ms\tremaining: 1.88s\n",
      "42:\tlearn: 0.6838950\ttotal: 517ms\tremaining: 1.89s\n",
      "43:\tlearn: 0.6838569\ttotal: 530ms\tremaining: 1.88s\n",
      "44:\tlearn: 0.6838419\ttotal: 538ms\tremaining: 1.85s\n",
      "45:\tlearn: 0.6837880\ttotal: 546ms\tremaining: 1.83s\n",
      "46:\tlearn: 0.6836604\ttotal: 562ms\tremaining: 1.83s\n",
      "47:\tlearn: 0.6836548\ttotal: 568ms\tremaining: 1.8s\n",
      "48:\tlearn: 0.6836028\ttotal: 584ms\tremaining: 1.8s\n",
      "49:\tlearn: 0.6834758\ttotal: 600ms\tremaining: 1.8s\n",
      "50:\tlearn: 0.6832493\ttotal: 616ms\tremaining: 1.8s\n",
      "51:\tlearn: 0.6830927\ttotal: 631ms\tremaining: 1.8s\n",
      "52:\tlearn: 0.6830465\ttotal: 640ms\tremaining: 1.77s\n",
      "53:\tlearn: 0.6828414\ttotal: 656ms\tremaining: 1.77s\n",
      "54:\tlearn: 0.6826903\ttotal: 672ms\tremaining: 1.77s\n",
      "55:\tlearn: 0.6825893\ttotal: 687ms\tremaining: 1.77s\n",
      "56:\tlearn: 0.6824653\ttotal: 704ms\tremaining: 1.76s\n",
      "57:\tlearn: 0.6822996\ttotal: 720ms\tremaining: 1.76s\n",
      "58:\tlearn: 0.6821113\ttotal: 737ms\tremaining: 1.76s\n",
      "59:\tlearn: 0.6819990\ttotal: 753ms\tremaining: 1.76s\n",
      "60:\tlearn: 0.6817919\ttotal: 769ms\tremaining: 1.75s\n",
      "61:\tlearn: 0.6816603\ttotal: 785ms\tremaining: 1.75s\n",
      "62:\tlearn: 0.6815000\ttotal: 801ms\tremaining: 1.74s\n",
      "63:\tlearn: 0.6814073\ttotal: 818ms\tremaining: 1.74s\n",
      "64:\tlearn: 0.6812921\ttotal: 835ms\tremaining: 1.73s\n",
      "65:\tlearn: 0.6811607\ttotal: 858ms\tremaining: 1.74s\n",
      "66:\tlearn: 0.6809812\ttotal: 877ms\tremaining: 1.74s\n",
      "67:\tlearn: 0.6808250\ttotal: 895ms\tremaining: 1.74s\n",
      "68:\tlearn: 0.6806460\ttotal: 912ms\tremaining: 1.73s\n",
      "69:\tlearn: 0.6804633\ttotal: 929ms\tremaining: 1.72s\n",
      "70:\tlearn: 0.6802471\ttotal: 945ms\tremaining: 1.72s\n",
      "71:\tlearn: 0.6801615\ttotal: 961ms\tremaining: 1.71s\n",
      "72:\tlearn: 0.6800161\ttotal: 977ms\tremaining: 1.7s\n",
      "73:\tlearn: 0.6798911\ttotal: 994ms\tremaining: 1.69s\n",
      "74:\tlearn: 0.6798266\ttotal: 1.01s\tremaining: 1.69s\n",
      "75:\tlearn: 0.6795964\ttotal: 1.03s\tremaining: 1.68s\n",
      "76:\tlearn: 0.6794943\ttotal: 1.04s\tremaining: 1.67s\n",
      "77:\tlearn: 0.6793150\ttotal: 1.06s\tremaining: 1.66s\n",
      "78:\tlearn: 0.6791340\ttotal: 1.08s\tremaining: 1.65s\n",
      "79:\tlearn: 0.6789688\ttotal: 1.09s\tremaining: 1.64s\n",
      "80:\tlearn: 0.6788715\ttotal: 1.11s\tremaining: 1.63s\n",
      "81:\tlearn: 0.6786931\ttotal: 1.13s\tremaining: 1.62s\n",
      "82:\tlearn: 0.6786009\ttotal: 1.14s\tremaining: 1.61s\n",
      "83:\tlearn: 0.6784056\ttotal: 1.16s\tremaining: 1.6s\n",
      "84:\tlearn: 0.6782664\ttotal: 1.17s\tremaining: 1.59s\n",
      "85:\tlearn: 0.6780465\ttotal: 1.19s\tremaining: 1.58s\n",
      "86:\tlearn: 0.6779340\ttotal: 1.21s\tremaining: 1.57s\n",
      "87:\tlearn: 0.6777499\ttotal: 1.22s\tremaining: 1.56s\n",
      "88:\tlearn: 0.6776678\ttotal: 1.24s\tremaining: 1.55s\n",
      "89:\tlearn: 0.6775801\ttotal: 1.26s\tremaining: 1.54s\n",
      "90:\tlearn: 0.6775257\ttotal: 1.27s\tremaining: 1.53s\n",
      "91:\tlearn: 0.6773990\ttotal: 1.29s\tremaining: 1.51s\n",
      "92:\tlearn: 0.6772745\ttotal: 1.31s\tremaining: 1.5s\n",
      "93:\tlearn: 0.6770699\ttotal: 1.32s\tremaining: 1.49s\n",
      "94:\tlearn: 0.6769313\ttotal: 1.34s\tremaining: 1.48s\n",
      "95:\tlearn: 0.6767531\ttotal: 1.36s\tremaining: 1.47s\n",
      "96:\tlearn: 0.6766714\ttotal: 1.37s\tremaining: 1.46s\n",
      "97:\tlearn: 0.6765430\ttotal: 1.39s\tremaining: 1.45s\n",
      "98:\tlearn: 0.6764027\ttotal: 1.41s\tremaining: 1.43s\n",
      "99:\tlearn: 0.6762961\ttotal: 1.42s\tremaining: 1.42s\n",
      "100:\tlearn: 0.6762168\ttotal: 1.44s\tremaining: 1.41s\n",
      "101:\tlearn: 0.6761214\ttotal: 1.46s\tremaining: 1.4s\n",
      "102:\tlearn: 0.6759947\ttotal: 1.47s\tremaining: 1.39s\n",
      "103:\tlearn: 0.6759018\ttotal: 1.49s\tremaining: 1.37s\n",
      "104:\tlearn: 0.6758039\ttotal: 1.5s\tremaining: 1.36s\n",
      "105:\tlearn: 0.6757046\ttotal: 1.52s\tremaining: 1.35s\n",
      "106:\tlearn: 0.6755396\ttotal: 1.53s\tremaining: 1.33s\n",
      "107:\tlearn: 0.6754432\ttotal: 1.55s\tremaining: 1.32s\n",
      "108:\tlearn: 0.6753341\ttotal: 1.57s\tremaining: 1.31s\n",
      "109:\tlearn: 0.6751679\ttotal: 1.58s\tremaining: 1.3s\n",
      "110:\tlearn: 0.6751171\ttotal: 1.6s\tremaining: 1.28s\n",
      "111:\tlearn: 0.6750740\ttotal: 1.62s\tremaining: 1.27s\n",
      "112:\tlearn: 0.6749486\ttotal: 1.63s\tremaining: 1.26s\n",
      "113:\tlearn: 0.6748768\ttotal: 1.65s\tremaining: 1.24s\n",
      "114:\tlearn: 0.6748017\ttotal: 1.67s\tremaining: 1.23s\n",
      "115:\tlearn: 0.6747268\ttotal: 1.68s\tremaining: 1.22s\n",
      "116:\tlearn: 0.6746348\ttotal: 1.7s\tremaining: 1.21s\n",
      "117:\tlearn: 0.6744802\ttotal: 1.72s\tremaining: 1.19s\n",
      "118:\tlearn: 0.6743617\ttotal: 1.74s\tremaining: 1.18s\n",
      "119:\tlearn: 0.6742674\ttotal: 1.75s\tremaining: 1.17s\n",
      "120:\tlearn: 0.6741708\ttotal: 1.77s\tremaining: 1.16s\n",
      "121:\tlearn: 0.6740376\ttotal: 1.79s\tremaining: 1.14s\n",
      "122:\tlearn: 0.6739536\ttotal: 1.81s\tremaining: 1.13s\n",
      "123:\tlearn: 0.6738101\ttotal: 1.83s\tremaining: 1.12s\n",
      "124:\tlearn: 0.6736755\ttotal: 1.84s\tremaining: 1.11s\n",
      "125:\tlearn: 0.6735690\ttotal: 1.86s\tremaining: 1.09s\n",
      "126:\tlearn: 0.6734481\ttotal: 1.89s\tremaining: 1.08s\n",
      "127:\tlearn: 0.6733058\ttotal: 1.9s\tremaining: 1.07s\n",
      "128:\tlearn: 0.6732420\ttotal: 1.92s\tremaining: 1.06s\n",
      "129:\tlearn: 0.6731583\ttotal: 1.94s\tremaining: 1.04s\n",
      "130:\tlearn: 0.6730673\ttotal: 1.95s\tremaining: 1.03s\n",
      "131:\tlearn: 0.6729756\ttotal: 1.97s\tremaining: 1.02s\n",
      "132:\tlearn: 0.6728961\ttotal: 1.99s\tremaining: 1s\n",
      "133:\tlearn: 0.6728279\ttotal: 2.01s\tremaining: 990ms\n",
      "134:\tlearn: 0.6727391\ttotal: 2.03s\tremaining: 976ms\n",
      "135:\tlearn: 0.6726406\ttotal: 2.04s\tremaining: 963ms\n",
      "136:\tlearn: 0.6725372\ttotal: 2.06s\tremaining: 950ms\n",
      "137:\tlearn: 0.6724300\ttotal: 2.08s\tremaining: 937ms\n",
      "138:\tlearn: 0.6723610\ttotal: 2.11s\tremaining: 924ms\n",
      "139:\tlearn: 0.6722327\ttotal: 2.12s\tremaining: 909ms\n",
      "140:\tlearn: 0.6721606\ttotal: 2.14s\tremaining: 895ms\n",
      "141:\tlearn: 0.6721076\ttotal: 2.15s\tremaining: 880ms\n",
      "142:\tlearn: 0.6720704\ttotal: 2.17s\tremaining: 866ms\n",
      "143:\tlearn: 0.6720243\ttotal: 2.19s\tremaining: 852ms\n",
      "144:\tlearn: 0.6719869\ttotal: 2.21s\tremaining: 837ms\n",
      "145:\tlearn: 0.6719178\ttotal: 2.22s\tremaining: 823ms\n",
      "146:\tlearn: 0.6718762\ttotal: 2.24s\tremaining: 808ms\n",
      "147:\tlearn: 0.6717947\ttotal: 2.26s\tremaining: 793ms\n",
      "148:\tlearn: 0.6716686\ttotal: 2.27s\tremaining: 779ms\n",
      "149:\tlearn: 0.6716514\ttotal: 2.29s\tremaining: 764ms\n",
      "150:\tlearn: 0.6715309\ttotal: 2.31s\tremaining: 749ms\n",
      "151:\tlearn: 0.6714685\ttotal: 2.32s\tremaining: 734ms\n",
      "152:\tlearn: 0.6713882\ttotal: 2.34s\tremaining: 719ms\n",
      "153:\tlearn: 0.6713063\ttotal: 2.36s\tremaining: 704ms\n",
      "154:\tlearn: 0.6712302\ttotal: 2.38s\tremaining: 690ms\n",
      "155:\tlearn: 0.6710994\ttotal: 2.39s\tremaining: 675ms\n",
      "156:\tlearn: 0.6710526\ttotal: 2.41s\tremaining: 660ms\n",
      "157:\tlearn: 0.6709074\ttotal: 2.42s\tremaining: 645ms\n",
      "158:\tlearn: 0.6707798\ttotal: 2.44s\tremaining: 630ms\n",
      "159:\tlearn: 0.6707430\ttotal: 2.46s\tremaining: 614ms\n",
      "160:\tlearn: 0.6706759\ttotal: 2.47s\tremaining: 599ms\n",
      "161:\tlearn: 0.6706010\ttotal: 2.49s\tremaining: 584ms\n",
      "162:\tlearn: 0.6704551\ttotal: 2.51s\tremaining: 569ms\n",
      "163:\tlearn: 0.6703437\ttotal: 2.52s\tremaining: 554ms\n",
      "164:\tlearn: 0.6702937\ttotal: 2.54s\tremaining: 539ms\n",
      "165:\tlearn: 0.6702573\ttotal: 2.56s\tremaining: 524ms\n",
      "166:\tlearn: 0.6701381\ttotal: 2.57s\tremaining: 509ms\n",
      "167:\tlearn: 0.6700971\ttotal: 2.59s\tremaining: 493ms\n",
      "168:\tlearn: 0.6700465\ttotal: 2.61s\tremaining: 478ms\n",
      "169:\tlearn: 0.6700167\ttotal: 2.62s\tremaining: 463ms\n",
      "170:\tlearn: 0.6699107\ttotal: 2.64s\tremaining: 448ms\n",
      "171:\tlearn: 0.6698829\ttotal: 2.66s\tremaining: 433ms\n",
      "172:\tlearn: 0.6698155\ttotal: 2.67s\tremaining: 418ms\n",
      "173:\tlearn: 0.6697024\ttotal: 2.69s\tremaining: 402ms\n",
      "174:\tlearn: 0.6695923\ttotal: 2.71s\tremaining: 387ms\n",
      "175:\tlearn: 0.6695016\ttotal: 2.72s\tremaining: 372ms\n",
      "176:\tlearn: 0.6694571\ttotal: 2.74s\tremaining: 356ms\n",
      "177:\tlearn: 0.6693928\ttotal: 2.76s\tremaining: 341ms\n",
      "178:\tlearn: 0.6692699\ttotal: 2.77s\tremaining: 325ms\n",
      "179:\tlearn: 0.6691900\ttotal: 2.79s\tremaining: 310ms\n",
      "180:\tlearn: 0.6691307\ttotal: 2.81s\tremaining: 295ms\n",
      "181:\tlearn: 0.6690973\ttotal: 2.82s\tremaining: 279ms\n",
      "182:\tlearn: 0.6689777\ttotal: 2.84s\tremaining: 264ms\n",
      "183:\tlearn: 0.6689294\ttotal: 2.85s\tremaining: 248ms\n",
      "184:\tlearn: 0.6688743\ttotal: 2.87s\tremaining: 233ms\n",
      "185:\tlearn: 0.6688173\ttotal: 2.89s\tremaining: 217ms\n",
      "186:\tlearn: 0.6687847\ttotal: 2.9s\tremaining: 202ms\n",
      "187:\tlearn: 0.6687193\ttotal: 2.92s\tremaining: 186ms\n",
      "188:\tlearn: 0.6686836\ttotal: 2.94s\tremaining: 171ms\n",
      "189:\tlearn: 0.6686220\ttotal: 2.95s\tremaining: 155ms\n",
      "190:\tlearn: 0.6685632\ttotal: 2.97s\tremaining: 140ms\n",
      "191:\tlearn: 0.6684895\ttotal: 2.98s\tremaining: 124ms\n",
      "192:\tlearn: 0.6683938\ttotal: 3s\tremaining: 109ms\n",
      "193:\tlearn: 0.6683317\ttotal: 3.02s\tremaining: 93.4ms\n",
      "194:\tlearn: 0.6682869\ttotal: 3.04s\tremaining: 77.8ms\n",
      "195:\tlearn: 0.6682406\ttotal: 3.05s\tremaining: 62.3ms\n",
      "196:\tlearn: 0.6681991\ttotal: 3.07s\tremaining: 46.8ms\n",
      "197:\tlearn: 0.6681182\ttotal: 3.09s\tremaining: 31.2ms\n",
      "198:\tlearn: 0.6680750\ttotal: 3.1s\tremaining: 15.6ms\n",
      "199:\tlearn: 0.6680104\ttotal: 3.12s\tremaining: 0us\n",
      "Learning rate set to 0.212981\n",
      "0:\tlearn: 0.6916807\ttotal: 8.31ms\tremaining: 1.65s\n",
      "1:\tlearn: 0.6904250\ttotal: 25.4ms\tremaining: 2.52s\n",
      "2:\tlearn: 0.6895454\ttotal: 35.8ms\tremaining: 2.35s\n",
      "3:\tlearn: 0.6892558\ttotal: 42ms\tremaining: 2.06s\n",
      "4:\tlearn: 0.6886495\ttotal: 58.8ms\tremaining: 2.29s\n",
      "5:\tlearn: 0.6883041\ttotal: 67.5ms\tremaining: 2.18s\n",
      "6:\tlearn: 0.6881052\ttotal: 76ms\tremaining: 2.1s\n",
      "7:\tlearn: 0.6877933\ttotal: 92.2ms\tremaining: 2.21s\n",
      "8:\tlearn: 0.6876518\ttotal: 100ms\tremaining: 2.13s\n",
      "9:\tlearn: 0.6875335\ttotal: 111ms\tremaining: 2.1s\n",
      "10:\tlearn: 0.6873541\ttotal: 127ms\tremaining: 2.19s\n",
      "11:\tlearn: 0.6872450\ttotal: 136ms\tremaining: 2.13s\n",
      "12:\tlearn: 0.6869597\ttotal: 152ms\tremaining: 2.18s\n",
      "13:\tlearn: 0.6868329\ttotal: 163ms\tremaining: 2.16s\n",
      "14:\tlearn: 0.6867171\ttotal: 186ms\tremaining: 2.3s\n",
      "15:\tlearn: 0.6866636\ttotal: 198ms\tremaining: 2.28s\n",
      "16:\tlearn: 0.6865226\ttotal: 218ms\tremaining: 2.35s\n",
      "17:\tlearn: 0.6864492\ttotal: 231ms\tremaining: 2.34s\n",
      "18:\tlearn: 0.6862961\ttotal: 253ms\tremaining: 2.41s\n",
      "19:\tlearn: 0.6862766\ttotal: 261ms\tremaining: 2.35s\n",
      "20:\tlearn: 0.6862688\ttotal: 268ms\tremaining: 2.29s\n",
      "21:\tlearn: 0.6860929\ttotal: 283ms\tremaining: 2.29s\n",
      "22:\tlearn: 0.6859611\ttotal: 300ms\tremaining: 2.31s\n",
      "23:\tlearn: 0.6859614\ttotal: 304ms\tremaining: 2.23s\n",
      "24:\tlearn: 0.6858491\ttotal: 321ms\tremaining: 2.25s\n",
      "25:\tlearn: 0.6858421\ttotal: 331ms\tremaining: 2.21s\n",
      "26:\tlearn: 0.6857233\ttotal: 348ms\tremaining: 2.23s\n",
      "27:\tlearn: 0.6857170\ttotal: 355ms\tremaining: 2.18s\n",
      "28:\tlearn: 0.6856764\ttotal: 363ms\tremaining: 2.14s\n",
      "29:\tlearn: 0.6855990\ttotal: 373ms\tremaining: 2.11s\n",
      "30:\tlearn: 0.6854446\ttotal: 389ms\tremaining: 2.12s\n",
      "31:\tlearn: 0.6853071\ttotal: 405ms\tremaining: 2.13s\n",
      "32:\tlearn: 0.6851335\ttotal: 421ms\tremaining: 2.13s\n",
      "33:\tlearn: 0.6850881\ttotal: 428ms\tremaining: 2.09s\n",
      "34:\tlearn: 0.6849071\ttotal: 444ms\tremaining: 2.09s\n",
      "35:\tlearn: 0.6848522\ttotal: 451ms\tremaining: 2.05s\n",
      "36:\tlearn: 0.6847753\ttotal: 464ms\tremaining: 2.04s\n",
      "37:\tlearn: 0.6847439\ttotal: 472ms\tremaining: 2.01s\n",
      "38:\tlearn: 0.6846653\ttotal: 478ms\tremaining: 1.97s\n",
      "39:\tlearn: 0.6845570\ttotal: 494ms\tremaining: 1.97s\n",
      "40:\tlearn: 0.6844580\ttotal: 509ms\tremaining: 1.97s\n",
      "41:\tlearn: 0.6844531\ttotal: 516ms\tremaining: 1.94s\n",
      "42:\tlearn: 0.6843754\ttotal: 532ms\tremaining: 1.94s\n",
      "43:\tlearn: 0.6842780\ttotal: 548ms\tremaining: 1.94s\n",
      "44:\tlearn: 0.6841139\ttotal: 564ms\tremaining: 1.94s\n",
      "45:\tlearn: 0.6841116\ttotal: 572ms\tremaining: 1.91s\n",
      "46:\tlearn: 0.6839252\ttotal: 589ms\tremaining: 1.92s\n",
      "47:\tlearn: 0.6837661\ttotal: 605ms\tremaining: 1.92s\n",
      "48:\tlearn: 0.6836247\ttotal: 620ms\tremaining: 1.91s\n",
      "49:\tlearn: 0.6834826\ttotal: 636ms\tremaining: 1.91s\n",
      "50:\tlearn: 0.6833528\ttotal: 653ms\tremaining: 1.91s\n",
      "51:\tlearn: 0.6832904\ttotal: 669ms\tremaining: 1.9s\n",
      "52:\tlearn: 0.6831975\ttotal: 685ms\tremaining: 1.9s\n",
      "53:\tlearn: 0.6830321\ttotal: 701ms\tremaining: 1.9s\n",
      "54:\tlearn: 0.6829055\ttotal: 717ms\tremaining: 1.89s\n",
      "55:\tlearn: 0.6827738\ttotal: 732ms\tremaining: 1.88s\n",
      "56:\tlearn: 0.6826021\ttotal: 747ms\tremaining: 1.88s\n",
      "57:\tlearn: 0.6824559\ttotal: 764ms\tremaining: 1.87s\n",
      "58:\tlearn: 0.6822799\ttotal: 780ms\tremaining: 1.86s\n",
      "59:\tlearn: 0.6821365\ttotal: 796ms\tremaining: 1.86s\n",
      "60:\tlearn: 0.6819650\ttotal: 812ms\tremaining: 1.85s\n",
      "61:\tlearn: 0.6817901\ttotal: 827ms\tremaining: 1.84s\n",
      "62:\tlearn: 0.6816801\ttotal: 844ms\tremaining: 1.83s\n",
      "63:\tlearn: 0.6814832\ttotal: 860ms\tremaining: 1.83s\n",
      "64:\tlearn: 0.6813076\ttotal: 876ms\tremaining: 1.82s\n",
      "65:\tlearn: 0.6811234\ttotal: 891ms\tremaining: 1.81s\n",
      "66:\tlearn: 0.6809569\ttotal: 906ms\tremaining: 1.8s\n",
      "67:\tlearn: 0.6807624\ttotal: 922ms\tremaining: 1.79s\n",
      "68:\tlearn: 0.6805546\ttotal: 937ms\tremaining: 1.78s\n",
      "69:\tlearn: 0.6804686\ttotal: 953ms\tremaining: 1.77s\n",
      "70:\tlearn: 0.6803153\ttotal: 969ms\tremaining: 1.76s\n",
      "71:\tlearn: 0.6801463\ttotal: 985ms\tremaining: 1.75s\n",
      "72:\tlearn: 0.6800191\ttotal: 1000ms\tremaining: 1.74s\n",
      "73:\tlearn: 0.6799449\ttotal: 1.01s\tremaining: 1.73s\n",
      "74:\tlearn: 0.6798016\ttotal: 1.03s\tremaining: 1.72s\n",
      "75:\tlearn: 0.6795754\ttotal: 1.05s\tremaining: 1.71s\n",
      "76:\tlearn: 0.6794123\ttotal: 1.06s\tremaining: 1.7s\n",
      "77:\tlearn: 0.6792084\ttotal: 1.08s\tremaining: 1.69s\n",
      "78:\tlearn: 0.6790893\ttotal: 1.1s\tremaining: 1.68s\n",
      "79:\tlearn: 0.6789926\ttotal: 1.11s\tremaining: 1.67s\n",
      "80:\tlearn: 0.6787813\ttotal: 1.13s\tremaining: 1.66s\n",
      "81:\tlearn: 0.6786171\ttotal: 1.15s\tremaining: 1.65s\n",
      "82:\tlearn: 0.6785010\ttotal: 1.16s\tremaining: 1.64s\n",
      "83:\tlearn: 0.6783887\ttotal: 1.18s\tremaining: 1.63s\n",
      "84:\tlearn: 0.6783285\ttotal: 1.2s\tremaining: 1.62s\n",
      "85:\tlearn: 0.6782020\ttotal: 1.21s\tremaining: 1.6s\n",
      "86:\tlearn: 0.6780416\ttotal: 1.23s\tremaining: 1.59s\n",
      "87:\tlearn: 0.6779723\ttotal: 1.24s\tremaining: 1.58s\n",
      "88:\tlearn: 0.6778798\ttotal: 1.26s\tremaining: 1.57s\n",
      "89:\tlearn: 0.6776994\ttotal: 1.27s\tremaining: 1.56s\n",
      "90:\tlearn: 0.6776225\ttotal: 1.29s\tremaining: 1.55s\n",
      "91:\tlearn: 0.6774113\ttotal: 1.31s\tremaining: 1.53s\n",
      "92:\tlearn: 0.6773056\ttotal: 1.32s\tremaining: 1.52s\n",
      "93:\tlearn: 0.6772096\ttotal: 1.34s\tremaining: 1.51s\n",
      "94:\tlearn: 0.6770601\ttotal: 1.36s\tremaining: 1.5s\n",
      "95:\tlearn: 0.6769629\ttotal: 1.37s\tremaining: 1.49s\n",
      "96:\tlearn: 0.6768511\ttotal: 1.39s\tremaining: 1.48s\n",
      "97:\tlearn: 0.6767509\ttotal: 1.41s\tremaining: 1.46s\n",
      "98:\tlearn: 0.6766375\ttotal: 1.42s\tremaining: 1.45s\n",
      "99:\tlearn: 0.6764968\ttotal: 1.44s\tremaining: 1.44s\n",
      "100:\tlearn: 0.6763376\ttotal: 1.46s\tremaining: 1.43s\n",
      "101:\tlearn: 0.6762607\ttotal: 1.47s\tremaining: 1.41s\n",
      "102:\tlearn: 0.6761243\ttotal: 1.49s\tremaining: 1.4s\n",
      "103:\tlearn: 0.6760082\ttotal: 1.5s\tremaining: 1.39s\n",
      "104:\tlearn: 0.6759139\ttotal: 1.52s\tremaining: 1.37s\n",
      "105:\tlearn: 0.6757871\ttotal: 1.53s\tremaining: 1.36s\n",
      "106:\tlearn: 0.6757292\ttotal: 1.55s\tremaining: 1.34s\n",
      "107:\tlearn: 0.6756117\ttotal: 1.56s\tremaining: 1.33s\n",
      "108:\tlearn: 0.6754975\ttotal: 1.58s\tremaining: 1.32s\n",
      "109:\tlearn: 0.6753833\ttotal: 1.59s\tremaining: 1.3s\n",
      "110:\tlearn: 0.6752603\ttotal: 1.61s\tremaining: 1.29s\n",
      "111:\tlearn: 0.6752201\ttotal: 1.62s\tremaining: 1.27s\n",
      "112:\tlearn: 0.6751369\ttotal: 1.64s\tremaining: 1.26s\n",
      "113:\tlearn: 0.6750798\ttotal: 1.66s\tremaining: 1.25s\n",
      "114:\tlearn: 0.6749396\ttotal: 1.67s\tremaining: 1.24s\n",
      "115:\tlearn: 0.6748298\ttotal: 1.69s\tremaining: 1.22s\n",
      "116:\tlearn: 0.6747328\ttotal: 1.7s\tremaining: 1.21s\n",
      "117:\tlearn: 0.6746559\ttotal: 1.72s\tremaining: 1.19s\n",
      "118:\tlearn: 0.6745690\ttotal: 1.73s\tremaining: 1.18s\n",
      "119:\tlearn: 0.6744867\ttotal: 1.75s\tremaining: 1.17s\n",
      "120:\tlearn: 0.6743881\ttotal: 1.77s\tremaining: 1.15s\n",
      "121:\tlearn: 0.6742637\ttotal: 1.78s\tremaining: 1.14s\n",
      "122:\tlearn: 0.6741492\ttotal: 1.8s\tremaining: 1.13s\n",
      "123:\tlearn: 0.6740672\ttotal: 1.82s\tremaining: 1.11s\n",
      "124:\tlearn: 0.6739717\ttotal: 1.83s\tremaining: 1.1s\n",
      "125:\tlearn: 0.6738247\ttotal: 1.85s\tremaining: 1.09s\n",
      "126:\tlearn: 0.6737778\ttotal: 1.87s\tremaining: 1.07s\n",
      "127:\tlearn: 0.6737086\ttotal: 1.89s\tremaining: 1.06s\n",
      "128:\tlearn: 0.6735841\ttotal: 1.9s\tremaining: 1.05s\n",
      "129:\tlearn: 0.6735155\ttotal: 1.92s\tremaining: 1.03s\n",
      "130:\tlearn: 0.6734108\ttotal: 1.94s\tremaining: 1.02s\n",
      "131:\tlearn: 0.6733160\ttotal: 1.96s\tremaining: 1.01s\n",
      "132:\tlearn: 0.6732189\ttotal: 1.97s\tremaining: 993ms\n",
      "133:\tlearn: 0.6730681\ttotal: 1.99s\tremaining: 979ms\n",
      "134:\tlearn: 0.6729937\ttotal: 2s\tremaining: 965ms\n",
      "135:\tlearn: 0.6728575\ttotal: 2.02s\tremaining: 951ms\n",
      "136:\tlearn: 0.6727579\ttotal: 2.04s\tremaining: 939ms\n",
      "137:\tlearn: 0.6726262\ttotal: 2.06s\tremaining: 924ms\n",
      "138:\tlearn: 0.6725795\ttotal: 2.07s\tremaining: 910ms\n",
      "139:\tlearn: 0.6725393\ttotal: 2.09s\tremaining: 895ms\n",
      "140:\tlearn: 0.6724650\ttotal: 2.11s\tremaining: 882ms\n",
      "141:\tlearn: 0.6723939\ttotal: 2.12s\tremaining: 867ms\n",
      "142:\tlearn: 0.6722851\ttotal: 2.14s\tremaining: 853ms\n",
      "143:\tlearn: 0.6722075\ttotal: 2.16s\tremaining: 839ms\n",
      "144:\tlearn: 0.6721603\ttotal: 2.19s\tremaining: 830ms\n",
      "145:\tlearn: 0.6720948\ttotal: 2.21s\tremaining: 816ms\n",
      "146:\tlearn: 0.6719545\ttotal: 2.22s\tremaining: 801ms\n",
      "147:\tlearn: 0.6719082\ttotal: 2.24s\tremaining: 786ms\n",
      "148:\tlearn: 0.6718046\ttotal: 2.25s\tremaining: 771ms\n",
      "149:\tlearn: 0.6717209\ttotal: 2.27s\tremaining: 756ms\n",
      "150:\tlearn: 0.6716162\ttotal: 2.29s\tremaining: 742ms\n",
      "151:\tlearn: 0.6715071\ttotal: 2.3s\tremaining: 727ms\n",
      "152:\tlearn: 0.6713662\ttotal: 2.32s\tremaining: 712ms\n",
      "153:\tlearn: 0.6712562\ttotal: 2.33s\tremaining: 697ms\n",
      "154:\tlearn: 0.6711709\ttotal: 2.35s\tremaining: 682ms\n",
      "155:\tlearn: 0.6710914\ttotal: 2.37s\tremaining: 667ms\n",
      "156:\tlearn: 0.6710211\ttotal: 2.38s\tremaining: 652ms\n",
      "157:\tlearn: 0.6709066\ttotal: 2.4s\tremaining: 637ms\n",
      "158:\tlearn: 0.6708697\ttotal: 2.41s\tremaining: 623ms\n",
      "159:\tlearn: 0.6708453\ttotal: 2.43s\tremaining: 607ms\n",
      "160:\tlearn: 0.6707878\ttotal: 2.44s\tremaining: 592ms\n",
      "161:\tlearn: 0.6707058\ttotal: 2.46s\tremaining: 577ms\n",
      "162:\tlearn: 0.6705665\ttotal: 2.48s\tremaining: 562ms\n",
      "163:\tlearn: 0.6705191\ttotal: 2.49s\tremaining: 547ms\n",
      "164:\tlearn: 0.6704181\ttotal: 2.51s\tremaining: 532ms\n",
      "165:\tlearn: 0.6703269\ttotal: 2.52s\tremaining: 517ms\n",
      "166:\tlearn: 0.6703061\ttotal: 2.54s\tremaining: 502ms\n",
      "167:\tlearn: 0.6702520\ttotal: 2.55s\tremaining: 487ms\n",
      "168:\tlearn: 0.6701498\ttotal: 2.57s\tremaining: 471ms\n",
      "169:\tlearn: 0.6700969\ttotal: 2.59s\tremaining: 456ms\n",
      "170:\tlearn: 0.6700284\ttotal: 2.6s\tremaining: 441ms\n",
      "171:\tlearn: 0.6699520\ttotal: 2.62s\tremaining: 426ms\n",
      "172:\tlearn: 0.6699064\ttotal: 2.63s\tremaining: 411ms\n",
      "173:\tlearn: 0.6697979\ttotal: 2.65s\tremaining: 396ms\n",
      "174:\tlearn: 0.6697466\ttotal: 2.67s\tremaining: 381ms\n",
      "175:\tlearn: 0.6696677\ttotal: 2.68s\tremaining: 366ms\n",
      "176:\tlearn: 0.6696335\ttotal: 2.7s\tremaining: 351ms\n",
      "177:\tlearn: 0.6695464\ttotal: 2.72s\tremaining: 336ms\n",
      "178:\tlearn: 0.6694561\ttotal: 2.73s\tremaining: 321ms\n",
      "179:\tlearn: 0.6694119\ttotal: 2.76s\tremaining: 306ms\n",
      "180:\tlearn: 0.6693662\ttotal: 2.78s\tremaining: 292ms\n",
      "181:\tlearn: 0.6693255\ttotal: 2.79s\tremaining: 276ms\n",
      "182:\tlearn: 0.6692913\ttotal: 2.81s\tremaining: 261ms\n",
      "183:\tlearn: 0.6692579\ttotal: 2.83s\tremaining: 246ms\n",
      "184:\tlearn: 0.6692129\ttotal: 2.84s\tremaining: 231ms\n",
      "185:\tlearn: 0.6691661\ttotal: 2.86s\tremaining: 215ms\n",
      "186:\tlearn: 0.6691155\ttotal: 2.88s\tremaining: 200ms\n",
      "187:\tlearn: 0.6690479\ttotal: 2.89s\tremaining: 185ms\n",
      "188:\tlearn: 0.6690027\ttotal: 2.91s\tremaining: 169ms\n",
      "189:\tlearn: 0.6689347\ttotal: 2.92s\tremaining: 154ms\n",
      "190:\tlearn: 0.6688563\ttotal: 2.94s\tremaining: 139ms\n",
      "191:\tlearn: 0.6687938\ttotal: 2.96s\tremaining: 123ms\n",
      "192:\tlearn: 0.6687352\ttotal: 2.97s\tremaining: 108ms\n",
      "193:\tlearn: 0.6686472\ttotal: 2.99s\tremaining: 92.4ms\n",
      "194:\tlearn: 0.6686071\ttotal: 3s\tremaining: 77.1ms\n",
      "195:\tlearn: 0.6684550\ttotal: 3.02s\tremaining: 61.6ms\n",
      "196:\tlearn: 0.6684405\ttotal: 3.04s\tremaining: 46.2ms\n",
      "197:\tlearn: 0.6683893\ttotal: 3.05s\tremaining: 30.8ms\n",
      "198:\tlearn: 0.6683275\ttotal: 3.07s\tremaining: 15.4ms\n",
      "199:\tlearn: 0.6683142\ttotal: 3.08s\tremaining: 0us\n",
      "Learning rate set to 0.5\n",
      "0:\tlearn: 0.6900223\ttotal: 8.77ms\tremaining: 868ms\n",
      "1:\tlearn: 0.6890245\ttotal: 15.3ms\tremaining: 749ms\n",
      "2:\tlearn: 0.6883344\ttotal: 29.4ms\tremaining: 951ms\n",
      "3:\tlearn: 0.6879478\ttotal: 41ms\tremaining: 984ms\n",
      "4:\tlearn: 0.6879432\ttotal: 46.5ms\tremaining: 884ms\n",
      "5:\tlearn: 0.6879281\ttotal: 51.9ms\tremaining: 813ms\n",
      "6:\tlearn: 0.6877389\ttotal: 63.6ms\tremaining: 845ms\n",
      "7:\tlearn: 0.6875632\ttotal: 70ms\tremaining: 805ms\n",
      "8:\tlearn: 0.6872497\ttotal: 82.3ms\tremaining: 832ms\n",
      "9:\tlearn: 0.6870218\ttotal: 94.7ms\tremaining: 852ms\n",
      "10:\tlearn: 0.6870288\ttotal: 103ms\tremaining: 832ms\n",
      "11:\tlearn: 0.6869748\ttotal: 109ms\tremaining: 801ms\n",
      "12:\tlearn: 0.6869717\ttotal: 115ms\tremaining: 770ms\n",
      "13:\tlearn: 0.6868595\ttotal: 127ms\tremaining: 781ms\n",
      "14:\tlearn: 0.6867555\ttotal: 135ms\tremaining: 766ms\n",
      "15:\tlearn: 0.6866995\ttotal: 145ms\tremaining: 759ms\n",
      "16:\tlearn: 0.6866883\ttotal: 149ms\tremaining: 726ms\n",
      "17:\tlearn: 0.6866122\ttotal: 161ms\tremaining: 732ms\n",
      "18:\tlearn: 0.6865822\ttotal: 170ms\tremaining: 724ms\n",
      "19:\tlearn: 0.6864307\ttotal: 182ms\tremaining: 727ms\n",
      "20:\tlearn: 0.6864454\ttotal: 189ms\tremaining: 710ms\n",
      "21:\tlearn: 0.6863628\ttotal: 196ms\tremaining: 694ms\n",
      "22:\tlearn: 0.6862864\ttotal: 208ms\tremaining: 697ms\n",
      "23:\tlearn: 0.6861647\ttotal: 222ms\tremaining: 702ms\n",
      "24:\tlearn: 0.6860547\ttotal: 234ms\tremaining: 703ms\n",
      "25:\tlearn: 0.6858739\ttotal: 246ms\tremaining: 701ms\n",
      "26:\tlearn: 0.6856415\ttotal: 259ms\tremaining: 699ms\n",
      "27:\tlearn: 0.6855036\ttotal: 270ms\tremaining: 695ms\n",
      "28:\tlearn: 0.6853696\ttotal: 283ms\tremaining: 692ms\n",
      "29:\tlearn: 0.6852251\ttotal: 295ms\tremaining: 688ms\n",
      "30:\tlearn: 0.6850902\ttotal: 307ms\tremaining: 683ms\n",
      "31:\tlearn: 0.6849879\ttotal: 320ms\tremaining: 681ms\n",
      "32:\tlearn: 0.6848470\ttotal: 335ms\tremaining: 680ms\n",
      "33:\tlearn: 0.6846605\ttotal: 347ms\tremaining: 674ms\n",
      "34:\tlearn: 0.6845006\ttotal: 360ms\tremaining: 668ms\n",
      "35:\tlearn: 0.6844119\ttotal: 372ms\tremaining: 662ms\n",
      "36:\tlearn: 0.6843152\ttotal: 384ms\tremaining: 654ms\n",
      "37:\tlearn: 0.6841649\ttotal: 396ms\tremaining: 646ms\n",
      "38:\tlearn: 0.6841087\ttotal: 410ms\tremaining: 641ms\n",
      "39:\tlearn: 0.6839940\ttotal: 422ms\tremaining: 634ms\n",
      "40:\tlearn: 0.6838525\ttotal: 434ms\tremaining: 624ms\n",
      "41:\tlearn: 0.6837117\ttotal: 446ms\tremaining: 616ms\n",
      "42:\tlearn: 0.6836389\ttotal: 458ms\tremaining: 607ms\n",
      "43:\tlearn: 0.6834783\ttotal: 469ms\tremaining: 597ms\n",
      "44:\tlearn: 0.6834089\ttotal: 481ms\tremaining: 588ms\n",
      "45:\tlearn: 0.6833216\ttotal: 493ms\tremaining: 579ms\n",
      "46:\tlearn: 0.6832197\ttotal: 507ms\tremaining: 571ms\n",
      "47:\tlearn: 0.6831590\ttotal: 519ms\tremaining: 562ms\n",
      "48:\tlearn: 0.6830536\ttotal: 530ms\tremaining: 552ms\n",
      "49:\tlearn: 0.6830283\ttotal: 542ms\tremaining: 542ms\n",
      "50:\tlearn: 0.6829245\ttotal: 554ms\tremaining: 532ms\n",
      "51:\tlearn: 0.6828950\ttotal: 565ms\tremaining: 522ms\n",
      "52:\tlearn: 0.6828522\ttotal: 577ms\tremaining: 512ms\n",
      "53:\tlearn: 0.6827484\ttotal: 589ms\tremaining: 502ms\n",
      "54:\tlearn: 0.6826553\ttotal: 601ms\tremaining: 492ms\n",
      "55:\tlearn: 0.6825307\ttotal: 614ms\tremaining: 482ms\n",
      "56:\tlearn: 0.6824478\ttotal: 625ms\tremaining: 472ms\n",
      "57:\tlearn: 0.6824314\ttotal: 637ms\tremaining: 462ms\n",
      "58:\tlearn: 0.6823538\ttotal: 649ms\tremaining: 451ms\n",
      "59:\tlearn: 0.6822482\ttotal: 661ms\tremaining: 440ms\n",
      "60:\tlearn: 0.6821633\ttotal: 672ms\tremaining: 430ms\n",
      "61:\tlearn: 0.6821139\ttotal: 684ms\tremaining: 419ms\n",
      "62:\tlearn: 0.6820445\ttotal: 696ms\tremaining: 409ms\n",
      "63:\tlearn: 0.6819590\ttotal: 708ms\tremaining: 398ms\n",
      "64:\tlearn: 0.6817926\ttotal: 720ms\tremaining: 388ms\n",
      "65:\tlearn: 0.6817553\ttotal: 732ms\tremaining: 377ms\n",
      "66:\tlearn: 0.6817095\ttotal: 744ms\tremaining: 367ms\n",
      "67:\tlearn: 0.6815917\ttotal: 756ms\tremaining: 356ms\n",
      "68:\tlearn: 0.6815372\ttotal: 768ms\tremaining: 345ms\n",
      "69:\tlearn: 0.6814956\ttotal: 780ms\tremaining: 334ms\n",
      "70:\tlearn: 0.6814284\ttotal: 794ms\tremaining: 324ms\n",
      "71:\tlearn: 0.6814093\ttotal: 806ms\tremaining: 313ms\n",
      "72:\tlearn: 0.6813678\ttotal: 818ms\tremaining: 303ms\n",
      "73:\tlearn: 0.6813341\ttotal: 830ms\tremaining: 292ms\n",
      "74:\tlearn: 0.6812967\ttotal: 846ms\tremaining: 282ms\n",
      "75:\tlearn: 0.6811874\ttotal: 863ms\tremaining: 272ms\n",
      "76:\tlearn: 0.6810830\ttotal: 879ms\tremaining: 263ms\n",
      "77:\tlearn: 0.6810129\ttotal: 896ms\tremaining: 253ms\n",
      "78:\tlearn: 0.6809442\ttotal: 912ms\tremaining: 242ms\n",
      "79:\tlearn: 0.6809080\ttotal: 931ms\tremaining: 233ms\n",
      "80:\tlearn: 0.6808228\ttotal: 952ms\tremaining: 223ms\n",
      "81:\tlearn: 0.6807450\ttotal: 968ms\tremaining: 213ms\n",
      "82:\tlearn: 0.6806812\ttotal: 983ms\tremaining: 201ms\n",
      "83:\tlearn: 0.6806029\ttotal: 1s\tremaining: 190ms\n",
      "84:\tlearn: 0.6805749\ttotal: 1.02s\tremaining: 180ms\n",
      "85:\tlearn: 0.6805321\ttotal: 1.03s\tremaining: 168ms\n",
      "86:\tlearn: 0.6804795\ttotal: 1.04s\tremaining: 156ms\n",
      "87:\tlearn: 0.6804569\ttotal: 1.05s\tremaining: 144ms\n",
      "88:\tlearn: 0.6804147\ttotal: 1.07s\tremaining: 132ms\n",
      "89:\tlearn: 0.6803277\ttotal: 1.08s\tremaining: 120ms\n",
      "90:\tlearn: 0.6803114\ttotal: 1.09s\tremaining: 108ms\n",
      "91:\tlearn: 0.6802872\ttotal: 1.1s\tremaining: 96.2ms\n",
      "92:\tlearn: 0.6802049\ttotal: 1.12s\tremaining: 84.2ms\n",
      "93:\tlearn: 0.6801481\ttotal: 1.13s\tremaining: 72.2ms\n",
      "94:\tlearn: 0.6801075\ttotal: 1.14s\tremaining: 60.2ms\n",
      "95:\tlearn: 0.6800757\ttotal: 1.16s\tremaining: 48.2ms\n",
      "96:\tlearn: 0.6800016\ttotal: 1.17s\tremaining: 36.2ms\n",
      "97:\tlearn: 0.6799653\ttotal: 1.18s\tremaining: 24.1ms\n",
      "98:\tlearn: 0.6799125\ttotal: 1.19s\tremaining: 12.1ms\n",
      "99:\tlearn: 0.6798412\ttotal: 1.21s\tremaining: 0us\n",
      "train_scores_catboost: 0.5370173903574595\n",
      "test_scores_catboost: 0.5390678366875098\n"
     ]
    }
   ],
   "source": [
    "param = {\n",
    "    'classifier__n_estimators': [100, 200],\n",
    "    'classifier__max_depth': [5, 10]}\n",
    "\n",
    "grid_catboost = GridSearchCV(pipe_for_catboost, param, cv=2, scoring='f1', verbose=False)\n",
    "grid_catboost.fit(X_train, y_train)\n",
    "predict_train_catboost = grid_catboost.predict(X_train)\n",
    "predict_test_catboost = grid_catboost.predict(X_val)\n",
    "train_scores_catboost = grid_catboost.best_score_\n",
    "val_scores_catboost = f1_score(y_val, predict_test_catboost)\n",
    "print(f'train_scores_catboost: {train_scores_catboost}')\n",
    "print(f'test_scores_catboost: {val_scores_catboost}')\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "d08dce4c",
   "metadata": {},
   "source": [
    "Смотрим на размеры выборок"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "4b840b36",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    38352\n",
       "0    37600\n",
       "Name: at_fault, dtype: int64"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "d6a56031",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    32874\n",
       "0    32228\n",
       "Name: at_fault, dtype: int64"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.value_counts()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "e3a50963",
   "metadata": {},
   "source": [
    "Выводим медрики моделей"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "07aea7e9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Модель</th>\n",
       "      <th>Метрика</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>RF</td>\n",
       "      <td>0.586173</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>LGBM</td>\n",
       "      <td>0.534920</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CatBoost</td>\n",
       "      <td>0.539068</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Модель   Метрика\n",
       "0        RF  0.586173\n",
       "1      LGBM  0.534920\n",
       "2  CatBoost  0.539068"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "data = {\n",
    "    'Модель': ['RF', 'LGBM', 'CatBoost'],\n",
    "    'Метрика': [val_scores_1, val_scores_2, val_scores_catboost ]}\n",
    "df = pd.DataFrame(data)\n",
    "df\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "1f2be180",
   "metadata": {},
   "source": [
    "`Вывод:` как мы видим, самую лучшую метрику паказала модель Случайного Леса (RF)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "cb28635a",
   "metadata": {},
   "source": [
    "## Шаг 6. Проверьте лучшую модель в работе"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "45bed8df",
   "metadata": {},
   "source": [
    "Проверим в деле нашу лучшую модель - RF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "51892db8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'classifier__max_depth': 10, 'classifier__n_estimators': 100}"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_tree.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "ab0555a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipe_for_tree = Pipeline([\n",
    "    ('preprocessor', preprocessor_for_tree),\n",
    "    ('classifier', RandomForestClassifier(random_state=42,max_depth=5, n_estimators=200))\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "3b173ada",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score: 0.5911542220131022\n"
     ]
    }
   ],
   "source": [
    "model = pipe_for_tree\n",
    "model.fit(X_train,y_train)\n",
    "predict_final = model.predict(X_test)\n",
    "\n",
    "f1 = f1_score(y_test, predict_final)\n",
    "print(f\"F1 score: {f1}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "8a95fa9b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#  y_test и predict_final это истинные и предсказанные значения, соответственно\n",
    "cm = confusion_matrix(y_test, predict_final)\n",
    "\n",
    "# Вывод confusion matrix\n",
    "plt.figure(figsize=(12, 6))\n",
    "sns.heatmap(cm, annot=True, fmt=\"d\", cmap=\"Reds\")\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.xlabel(\"Predicted Labels\")\n",
    "plt.ylabel(\"True Labels\")\n",
    "plt.show()\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "1530f3cb",
   "metadata": {},
   "source": [
    "`Вывод:` Мы видим, что количество истинно-отрицательных показаний (True-negative) самое высокое, при этом  истинно-положительные (True-Positive) на треьем месте.\n",
    "\n",
    "Это говорит о хорошей предсказательной способности модели в случае негативных значений."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "a0cc4f72",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.49683640567013443"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_test,predict_final)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "66906879",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7296729807005004"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(y_test,predict_final)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "adceb1a4",
   "metadata": {},
   "source": [
    "`Вывод:` лучшая модель Случайного Леса показала \n",
    "\n",
    "Полнота: 0.5029506600961245<br>\n",
    "Точность: 0.7150765504714125\n",
    "\n",
    "Полнота (recall) измеряет долю правильно классифицированных положительных случаев (true positives) относительно общего количества фактических положительных случаев. В данном случае полнота на тестовой выборке составляет 0.5029506600961245. Это означает, что модель правильно классифицировала 50% всех фактических положительных случаев.\n",
    "\n",
    "Точность (precision) измеряет долю правильно классифицированных положительных случаев (true positives) относительно всех случаев, которые модель предсказала как положительные (true positives + false positives). В данном случае точность на тестовой выборке составляет 0.7150765504714125. Это означает, что из всех случаев, которые модель предсказала как положительные, 71% действительно являются положительными."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "89069db9",
   "metadata": {},
   "source": [
    "## Проведите анализ важности факторов ДТП"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a958cd4c",
   "metadata": {},
   "source": [
    "Выводим график для наглядности"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "894cc511",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance = model.named_steps['classifier'].feature_importances_\n",
    "\n",
    "\n",
    "# Создаём словарь с именами параметров и их важностью\n",
    "feature_importance_dict = dict(zip(preprocessor_for_tree.transformers_[0][1].get_feature_names_out(), importance))\n",
    "\n",
    "\n",
    "# Сортируем важность параметров по убыванию\n",
    "sorted_feature_importance = sorted(feature_importance_dict.items(), key=lambda x: x[1], reverse=True)\n",
    "features = [feature for feature, _ in sorted_feature_importance]\n",
    "importance = [importance for _, importance in sorted_feature_importance]\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.bar(features, importance)\n",
    "plt.xticks(rotation=90)\n",
    "plt.xlabel(\"Features\")\n",
    "plt.ylabel(\"Importance\")\n",
    "plt.title(\"Feature Importance\")\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "0c1f3980",
   "metadata": {},
   "source": [
    "Проведём исследование признака 'party_sobriety' и зависимость целевого параметра."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "5801d6a7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>at_fault</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>party_sobriety</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>had been drinking, impairment unknown</th>\n",
       "      <td>278</td>\n",
       "      <td>1428</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>had been drinking, not under influence</th>\n",
       "      <td>461</td>\n",
       "      <td>1067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>had been drinking, under influence</th>\n",
       "      <td>507</td>\n",
       "      <td>10692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>had not been drinking</th>\n",
       "      <td>101804</td>\n",
       "      <td>78770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>impairment unknown</th>\n",
       "      <td>3550</td>\n",
       "      <td>16838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>not applicable</th>\n",
       "      <td>828</td>\n",
       "      <td>783</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "at_fault                                     0      1\n",
       "party_sobriety                                       \n",
       "had been drinking, impairment unknown      278   1428\n",
       "had been drinking, not under influence     461   1067\n",
       "had been drinking, under influence         507  10692\n",
       "had not been drinking                   101804  78770\n",
       "impairment unknown                        3550  16838\n",
       "not applicable                             828    783"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "union_df.pivot_table(index='party_sobriety', columns='at_fault', values='distance', aggfunc='count')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "92e09c8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "union_df.pivot_table(index='party_sobriety', columns='at_fault', values='distance', aggfunc='count').plot(kind='barh')\n",
    "\n",
    "plt.xticks(rotation=30)\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "9c7bd910",
   "metadata": {},
   "source": [
    "`Вывод: ` большая зависимость ДТП по неизвестной причине impairment unknown (неизвестное ухудшение). Среди невыпивших, вероятность попадания / непопадания в ДТП отличается незначительно.\n",
    "\n",
    "Саямая явная зависимость - между ДТП и выпившим водителем под алкогольным влиянием."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "299b10ed",
   "metadata": {},
   "source": [
    "`Общий вывод: `\n",
    "\n",
    "**В результате проведённого исследования зависимости риска ДТП от разных дорожых факторов, были выполенны следущие этапы:**\n",
    "\n",
    "- Импорт из базы данных при помощи SQL \n",
    "- Ознакосмление с данными о ДТП из трёх разных таблиц vehicles, collisions и parties.\n",
    "- был провдён статистический анализ данных ДТП который выявил следущее:\n",
    "подавляющее большинство ДТП происходит с января по май. предположительно, это связанно с суровыми погодными условиями зимой и весенней распутицей, а также из-за переходного периода, смены зимних шин на летние (несвоевременная смена шин).\n",
    "- для потдверждения нашего предположения мы разделили задачу на подзадачи. \n",
    "\n",
    "- Изучить зависимость ДТП от состояния дорожного полотна\n",
    "- Изучить зависимость ДТП от состояния освящённости дороги\n",
    "- Изучить зависимость ДТП при нахождении на перекрёстке дорог\n",
    "- Изучить есть зависимость ДТП от уровня опасности участка дорого\n",
    "- Изучить зависимость ДТП от погодных условий\n",
    "- Изучить зависимость ДТП от использования громкой связи за рулём\n",
    "- Изучить зависимость ДТП от времени суток\n",
    "\n",
    "Мы выборочно проанализировали боллее подробно две из них: зависимость ДТП от погодных условий и зависимость ДТП от состояния освящённости дороги.\n",
    "\n",
    "Из проведённого анализа следует, что за исключением параметра 'ясная погода' сколько нибудь значительной разницы между присутствием или отсутствием погодного явления не надлюдается.\n",
    "\n",
    "ДТП проиходят одинаково часто в дождь или при его отсутствии.\n",
    "\n",
    "Папаметр clear - (ясная погода) показывает незначительную связь, на грани погрешности.\n",
    "\n",
    "Также выявлено, что не прослеживается корреляции отсещённости и ДТП, за исключением наличия / отсутствия дневного света, но эти показатели очень близки и поэтому, ненадёжны. Можно сделать до вывод что явной взаимосвязи нет.\n",
    "\n",
    "**Для предсказания вероятности ДТП были созданы и обучены модели RF, LGBM, Catboost в результате, лучшей моделью оказалась RF**\n",
    "\n",
    "в результате применения лучшей модели RF на тестовой выборке были достигнуты следущие показания метрики\n",
    "\n",
    "Полнота: 0.5029506600961245<br>\n",
    "Точность: 0.7150765504714125\n",
    "\n",
    "В итоге был определён фактор наиболее сильно врлияющий на колво ДТП - это опьянение, после подробного анализа выяснилось, что \n",
    "`саямая явная зависимость - между ДТП и выпившим водителем кторый находится под алкогольным влиянием.`\n",
    "\n",
    "\n",
    "`В качестве рекомендации для дальнейшего развития проекта можно предложить следущее:`\n",
    "Принимая во внимание что пьянство за рулём - основной фактор влияющий на количество ДТП, имеет смысл ввести следующие нововведения для водителей:\n",
    "- применение встроенных алкотестеров\n",
    "- системы распознавания водителя, чтобы убедиться что тест проводит нужный человек. Можно использовать скан радушки или лица."
   ]
  }
 ],
 "metadata": {
  "ExecuteTimeLog": [
   {
    "duration": 2125,
    "start_time": "2023-06-21T10:23:07.045Z"
   },
   {
    "duration": 3,
    "start_time": "2023-06-21T10:23:16.542Z"
   },
   {
    "duration": 8,
    "start_time": "2023-06-21T10:23:24.042Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T10:23:28.295Z"
   },
   {
    "duration": 21993,
    "start_time": "2023-06-21T10:23:42.396Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T10:24:22.105Z"
   },
   {
    "duration": 20442,
    "start_time": "2023-06-21T10:25:03.220Z"
   },
   {
    "duration": 39,
    "start_time": "2023-06-21T10:25:29.801Z"
   },
   {
    "duration": 36650,
    "start_time": "2023-06-21T10:26:05.140Z"
   },
   {
    "duration": 81,
    "start_time": "2023-06-21T10:28:33.774Z"
   },
   {
    "duration": 1958,
    "start_time": "2023-06-21T10:28:51.211Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T10:28:53.171Z"
   },
   {
    "duration": 20700,
    "start_time": "2023-06-21T10:28:53.802Z"
   },
   {
    "duration": 20078,
    "start_time": "2023-06-21T10:29:14.510Z"
   },
   {
    "duration": 4865,
    "start_time": "2023-06-21T10:29:37.935Z"
   },
   {
    "duration": 5761,
    "start_time": "2023-06-21T10:29:47.110Z"
   },
   {
    "duration": 63,
    "start_time": "2023-06-21T10:30:16.209Z"
   },
   {
    "duration": 2063,
    "start_time": "2023-06-21T10:30:19.447Z"
   },
   {
    "duration": 99,
    "start_time": "2023-06-21T10:30:24.104Z"
   },
   {
    "duration": 17,
    "start_time": "2023-06-21T10:30:26.490Z"
   },
   {
    "duration": 19142,
    "start_time": "2023-06-21T10:30:30.694Z"
   },
   {
    "duration": 23410,
    "start_time": "2023-06-21T10:31:17.031Z"
   },
   {
    "duration": 6326,
    "start_time": "2023-06-21T10:31:48.817Z"
   },
   {
    "duration": 6325,
    "start_time": "2023-06-21T10:31:57.215Z"
   },
   {
    "duration": 102,
    "start_time": "2023-06-21T10:32:27.731Z"
   },
   {
    "duration": 16,
    "start_time": "2023-06-21T10:32:31.897Z"
   },
   {
    "duration": 2126,
    "start_time": "2023-06-21T10:33:26.957Z"
   },
   {
    "duration": 20806,
    "start_time": "2023-06-21T10:33:29.085Z"
   },
   {
    "duration": 20049,
    "start_time": "2023-06-21T10:34:24.636Z"
   },
   {
    "duration": 45,
    "start_time": "2023-06-21T10:35:00.434Z"
   },
   {
    "duration": 23676,
    "start_time": "2023-06-21T10:35:32.337Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T10:36:00.318Z"
   },
   {
    "duration": 59,
    "start_time": "2023-06-21T10:36:05.780Z"
   },
   {
    "duration": 2724,
    "start_time": "2023-06-21T10:36:34.576Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T10:36:40.827Z"
   },
   {
    "duration": 43,
    "start_time": "2023-06-21T10:36:47.427Z"
   },
   {
    "duration": 21,
    "start_time": "2023-06-21T10:37:45.502Z"
   },
   {
    "duration": 738,
    "start_time": "2023-06-21T10:38:04.736Z"
   },
   {
    "duration": 3958,
    "start_time": "2023-06-21T10:38:08.451Z"
   },
   {
    "duration": 27,
    "start_time": "2023-06-21T10:38:50.543Z"
   },
   {
    "duration": 2719,
    "start_time": "2023-06-21T10:41:41.547Z"
   },
   {
    "duration": 6,
    "start_time": "2023-06-21T10:41:52.780Z"
   },
   {
    "duration": 3114,
    "start_time": "2023-06-21T10:42:04.397Z"
   },
   {
    "duration": 9,
    "start_time": "2023-06-21T10:42:09.150Z"
   },
   {
    "duration": 1614,
    "start_time": "2023-06-21T10:42:43.615Z"
   },
   {
    "duration": 16112,
    "start_time": "2023-06-21T10:42:45.231Z"
   },
   {
    "duration": 1498,
    "start_time": "2023-06-21T10:43:21.504Z"
   },
   {
    "duration": 14843,
    "start_time": "2023-06-21T10:43:23.007Z"
   },
   {
    "duration": 15163,
    "start_time": "2023-06-21T10:43:37.852Z"
   },
   {
    "duration": 19839,
    "start_time": "2023-06-21T10:44:00.391Z"
   },
   {
    "duration": 11410,
    "start_time": "2023-06-21T10:44:22.079Z"
   },
   {
    "duration": 28,
    "start_time": "2023-06-21T10:44:47.402Z"
   },
   {
    "duration": 4,
    "start_time": "2023-06-21T10:45:01.350Z"
   },
   {
    "duration": 1525,
    "start_time": "2023-06-21T10:46:44.806Z"
   },
   {
    "duration": 614,
    "start_time": "2023-06-21T10:48:45.232Z"
   },
   {
    "duration": 37832,
    "start_time": "2023-06-21T10:48:55.091Z"
   },
   {
    "duration": 130,
    "start_time": "2023-06-21T10:52:02.944Z"
   },
   {
    "duration": 137,
    "start_time": "2023-06-21T10:52:31.508Z"
   },
   {
    "duration": 111,
    "start_time": "2023-06-21T10:52:48.603Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T10:54:26.890Z"
   },
   {
    "duration": 195,
    "start_time": "2023-06-21T10:54:28.816Z"
   },
   {
    "duration": 15836,
    "start_time": "2023-06-21T10:54:38.903Z"
   },
   {
    "duration": 731,
    "start_time": "2023-06-21T10:54:57.914Z"
   },
   {
    "duration": 20546,
    "start_time": "2023-06-21T10:55:22.103Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-21T10:56:35.324Z"
   },
   {
    "duration": 136499,
    "start_time": "2023-06-21T10:56:38.715Z"
   },
   {
    "duration": 43,
    "start_time": "2023-06-21T10:59:08.944Z"
   },
   {
    "duration": 11,
    "start_time": "2023-06-21T10:59:31.707Z"
   },
   {
    "duration": 2499,
    "start_time": "2023-06-21T10:59:37.505Z"
   },
   {
    "duration": 3,
    "start_time": "2023-06-21T10:59:40.011Z"
   },
   {
    "duration": 11,
    "start_time": "2023-06-21T10:59:53.301Z"
   },
   {
    "duration": 12,
    "start_time": "2023-06-21T11:00:06.630Z"
   },
   {
    "duration": 442,
    "start_time": "2023-06-21T11:00:13.402Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-21T11:00:26.118Z"
   },
   {
    "duration": 6,
    "start_time": "2023-06-21T11:00:30.303Z"
   },
   {
    "duration": 23,
    "start_time": "2023-06-21T11:00:33.940Z"
   },
   {
    "duration": 1401,
    "start_time": "2023-06-22T12:18:12.064Z"
   },
   {
    "duration": 112,
    "start_time": "2023-06-22T12:18:45.475Z"
   },
   {
    "duration": 147,
    "start_time": "2023-06-22T12:19:57.851Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-22T12:19:58.252Z"
   },
   {
    "duration": 24235,
    "start_time": "2023-06-22T12:20:02.621Z"
   },
   {
    "duration": 37,
    "start_time": "2023-06-22T12:20:33.209Z"
   },
   {
    "duration": 11013,
    "start_time": "2023-06-22T12:21:36.015Z"
   },
   {
    "duration": 3533,
    "start_time": "2023-06-22T12:22:18.431Z"
   },
   {
    "duration": 22,
    "start_time": "2023-06-22T12:22:37.066Z"
   },
   {
    "duration": 3519,
    "start_time": "2023-06-22T12:22:43.635Z"
   },
   {
    "duration": 1349,
    "start_time": "2023-06-22T12:23:33.406Z"
   },
   {
    "duration": 23286,
    "start_time": "2023-06-22T12:23:34.757Z"
   },
   {
    "duration": 1393,
    "start_time": "2023-06-22T12:24:44.016Z"
   },
   {
    "duration": 22986,
    "start_time": "2023-06-22T12:24:46.736Z"
   },
   {
    "duration": 11663,
    "start_time": "2023-06-22T12:25:14.930Z"
   },
   {
    "duration": 3628,
    "start_time": "2023-06-22T12:25:29.050Z"
   },
   {
    "duration": 3688,
    "start_time": "2023-06-22T12:25:37.443Z"
   },
   {
    "duration": 4,
    "start_time": "2023-06-23T05:38:34.568Z"
   },
   {
    "duration": 65,
    "start_time": "2023-06-23T05:38:42.918Z"
   },
   {
    "duration": 2,
    "start_time": "2023-06-23T05:38:46.651Z"
   },
   {
    "duration": 3,
    "start_time": "2023-06-23T05:38:46.838Z"
   },
   {
    "duration": 8,
    "start_time": "2023-06-23T05:38:49.450Z"
   },
   {
    "duration": 8,
    "start_time": "2023-06-23T05:38:53.965Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-23T05:38:59.131Z"
   },
   {
    "duration": 10642,
    "start_time": "2023-06-23T05:39:01.715Z"
   },
   {
    "duration": 11633,
    "start_time": "2023-06-23T05:39:24.019Z"
   },
   {
    "duration": 45,
    "start_time": "2023-06-23T05:39:45.370Z"
   },
   {
    "duration": 1228,
    "start_time": "2023-06-23T05:40:25.393Z"
   },
   {
    "duration": 10358,
    "start_time": "2023-06-23T05:40:26.623Z"
   },
   {
    "duration": 9853,
    "start_time": "2023-06-23T05:40:40.589Z"
   },
   {
    "duration": 10660,
    "start_time": "2023-06-23T05:41:09.579Z"
   },
   {
    "duration": 6077,
    "start_time": "2023-06-23T05:41:20.240Z"
   },
   {
    "duration": 24,
    "start_time": "2023-06-23T05:41:26.319Z"
   },
   {
    "duration": 9858,
    "start_time": "2023-06-23T05:42:36.584Z"
   },
   {
    "duration": 30,
    "start_time": "2023-06-23T05:42:54.075Z"
   },
   {
    "duration": 253,
    "start_time": "2023-06-23T05:43:38.974Z"
   },
   {
    "duration": 1218,
    "start_time": "2023-06-23T05:43:59.291Z"
   },
   {
    "duration": 795,
    "start_time": "2023-06-23T05:44:11.042Z"
   },
   {
    "duration": 712,
    "start_time": "2023-06-23T05:44:28.214Z"
   },
   {
    "duration": 1176,
    "start_time": "2023-06-23T05:48:56.092Z"
   },
   {
    "duration": 11964,
    "start_time": "2023-06-23T05:48:58.369Z"
   },
   {
    "duration": 3,
    "start_time": "2023-06-23T05:49:13.519Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-23T05:49:13.928Z"
   },
   {
    "duration": 1679,
    "start_time": "2023-06-23T05:49:24.643Z"
   },
   {
    "duration": 29,
    "start_time": "2023-06-23T05:49:40.633Z"
   },
   {
    "duration": 103,
    "start_time": "2023-06-23T05:49:58.821Z"
   },
   {
    "duration": 37,
    "start_time": "2023-06-23T05:50:03.625Z"
   },
   {
    "duration": 631,
    "start_time": "2023-06-23T05:50:48.759Z"
   },
   {
    "duration": 35,
    "start_time": "2023-06-23T05:51:06.392Z"
   },
   {
    "duration": 1981,
    "start_time": "2023-06-28T06:44:55.210Z"
   },
   {
    "duration": 13879,
    "start_time": "2023-06-28T06:44:57.193Z"
   },
   {
    "duration": 2141,
    "start_time": "2023-06-28T06:45:11.074Z"
   },
   {
    "duration": 269,
    "start_time": "2023-06-28T06:45:13.218Z"
   },
   {
    "duration": 654,
    "start_time": "2023-06-28T06:45:13.489Z"
   },
   {
    "duration": 11011,
    "start_time": "2023-06-28T06:45:14.146Z"
   },
   {
    "duration": 400,
    "start_time": "2023-06-28T06:45:25.159Z"
   },
   {
    "duration": 1991,
    "start_time": "2023-06-28T06:45:25.567Z"
   },
   {
    "duration": 3496,
    "start_time": "2023-06-28T06:45:27.560Z"
   },
   {
    "duration": 4215,
    "start_time": "2023-06-28T06:45:31.060Z"
   },
   {
    "duration": 302,
    "start_time": "2023-06-28T06:45:35.277Z"
   },
   {
    "duration": 1180,
    "start_time": "2023-06-28T06:45:35.580Z"
   },
   {
    "duration": 150,
    "start_time": "2023-06-28T06:45:36.768Z"
   },
   {
    "duration": 152,
    "start_time": "2023-06-28T06:45:36.920Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.075Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.077Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.079Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.080Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.080Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.082Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.083Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.084Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.085Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.086Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.087Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.088Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.089Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.089Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.091Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.092Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.093Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.094Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.095Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.096Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.097Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.098Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.100Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.101Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.102Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.103Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.105Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.106Z"
   },
   {
    "duration": 0,
    "start_time": "2023-06-28T06:45:37.107Z"
   },
   {
    "duration": 1745,
    "start_time": "2023-06-28T10:25:10.830Z"
   },
   {
    "duration": 12612,
    "start_time": "2023-06-28T10:25:13.845Z"
   },
   {
    "duration": 4,
    "start_time": "2023-06-28T10:25:52.471Z"
   },
   {
    "duration": 260,
    "start_time": "2023-06-28T10:26:34.588Z"
   },
   {
    "duration": 134,
    "start_time": "2023-06-28T10:26:37.245Z"
   },
   {
    "duration": 96,
    "start_time": "2023-06-28T10:28:50.492Z"
   },
   {
    "duration": 152,
    "start_time": "2023-06-28T10:29:23.688Z"
   },
   {
    "duration": 12,
    "start_time": "2023-06-28T10:29:25.757Z"
   },
   {
    "duration": 27,
    "start_time": "2023-06-28T10:39:59.358Z"
   },
   {
    "duration": 231,
    "start_time": "2023-06-28T10:40:06.517Z"
   },
   {
    "duration": 157,
    "start_time": "2023-06-28T10:40:37.090Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-28T10:40:37.471Z"
   },
   {
    "duration": 634,
    "start_time": "2023-06-28T10:41:24.176Z"
   },
   {
    "duration": 21,
    "start_time": "2023-06-28T10:42:24.791Z"
   },
   {
    "duration": 123,
    "start_time": "2023-06-28T10:42:48.308Z"
   },
   {
    "duration": 160,
    "start_time": "2023-06-28T10:44:19.313Z"
   },
   {
    "duration": 6,
    "start_time": "2023-06-28T10:44:19.681Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-28T10:44:28.273Z"
   },
   {
    "duration": 12,
    "start_time": "2023-06-28T10:44:30.571Z"
   },
   {
    "duration": 156,
    "start_time": "2023-06-28T10:44:39.987Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-28T10:44:42.138Z"
   },
   {
    "duration": 7,
    "start_time": "2023-06-28T10:44:55.779Z"
   },
   {
    "duration": 656,
    "start_time": "2023-06-28T10:45:03.373Z"
   },
   {
    "duration": 4,
    "start_time": "2023-06-28T10:45:06.949Z"
   },
   {
    "duration": 15,
    "start_time": "2023-06-28T10:45:09.139Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-28T10:45:16.098Z"
   },
   {
    "duration": 8,
    "start_time": "2023-06-28T10:45:18.291Z"
   },
   {
    "duration": 4,
    "start_time": "2023-06-28T10:46:37.876Z"
   },
   {
    "duration": 18,
    "start_time": "2023-06-28T10:46:39.852Z"
   },
   {
    "duration": 115,
    "start_time": "2023-06-28T10:46:56.994Z"
   },
   {
    "duration": 4,
    "start_time": "2023-06-28T10:47:00.569Z"
   },
   {
    "duration": 15,
    "start_time": "2023-06-28T10:47:02.373Z"
   },
   {
    "duration": 193,
    "start_time": "2023-06-28T10:48:37.860Z"
   },
   {
    "duration": 124,
    "start_time": "2023-06-28T10:48:45.284Z"
   },
   {
    "duration": 124,
    "start_time": "2023-06-28T10:49:06.151Z"
   },
   {
    "duration": 140,
    "start_time": "2023-06-28T10:49:19.004Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-28T10:50:04.539Z"
   },
   {
    "duration": 165,
    "start_time": "2023-06-28T10:50:45.045Z"
   },
   {
    "duration": 128,
    "start_time": "2023-06-28T10:50:57.393Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-28T10:51:20.795Z"
   },
   {
    "duration": 183,
    "start_time": "2023-06-28T10:51:25.993Z"
   },
   {
    "duration": 157,
    "start_time": "2023-06-28T10:51:44.243Z"
   },
   {
    "duration": 182,
    "start_time": "2023-06-28T10:51:55.936Z"
   },
   {
    "duration": 308,
    "start_time": "2023-06-28T10:52:31.930Z"
   },
   {
    "duration": 358,
    "start_time": "2023-06-28T10:52:38.634Z"
   },
   {
    "duration": 226,
    "start_time": "2023-06-28T10:52:50.896Z"
   },
   {
    "duration": 251,
    "start_time": "2023-06-28T10:53:34.720Z"
   },
   {
    "duration": 231,
    "start_time": "2023-06-28T10:55:09.871Z"
   },
   {
    "duration": 227,
    "start_time": "2023-06-28T10:55:16.870Z"
   },
   {
    "duration": 232,
    "start_time": "2023-06-28T10:55:35.174Z"
   },
   {
    "duration": 245,
    "start_time": "2023-06-28T10:55:43.344Z"
   },
   {
    "duration": 238,
    "start_time": "2023-06-28T10:55:46.955Z"
   },
   {
    "duration": 258,
    "start_time": "2023-06-28T10:56:17.240Z"
   },
   {
    "duration": 163,
    "start_time": "2023-06-28T10:56:39.839Z"
   },
   {
    "duration": 13,
    "start_time": "2023-06-28T11:04:47.518Z"
   },
   {
    "duration": 48,
    "start_time": "2023-06-28T11:09:10.785Z"
   },
   {
    "duration": 1825,
    "start_time": "2023-06-28T11:09:13.696Z"
   },
   {
    "duration": 16846,
    "start_time": "2023-06-28T11:09:17.874Z"
   },
   {
    "duration": 13520,
    "start_time": "2023-06-28T11:12:55.828Z"
   },
   {
    "duration": 14,
    "start_time": "2023-06-28T11:13:13.436Z"
   },
   {
    "duration": 302,
    "start_time": "2023-06-28T11:15:28.745Z"
   },
   {
    "duration": 11,
    "start_time": "2023-06-28T11:16:00.311Z"
   },
   {
    "duration": 8657,
    "start_time": "2023-06-28T11:18:52.524Z"
   },
   {
    "duration": 14,
    "start_time": "2023-06-28T11:19:05.575Z"
   },
   {
    "duration": 350,
    "start_time": "2023-06-28T11:19:50.150Z"
   },
   {
    "duration": 2639,
    "start_time": "2023-06-28T11:20:00.943Z"
   },
   {
    "duration": 12,
    "start_time": "2023-06-28T11:20:06.316Z"
   },
   {
    "duration": 1781,
    "start_time": "2023-06-28T11:20:24.780Z"
   },
   {
    "duration": 13,
    "start_time": "2023-06-28T11:20:26.567Z"
   },
   {
    "duration": 361,
    "start_time": "2023-06-28T11:20:38.969Z"
   },
   {
    "duration": 338,
    "start_time": "2023-06-28T11:21:59.702Z"
   },
   {
    "duration": 379,
    "start_time": "2023-06-28T11:22:09.638Z"
   },
   {
    "duration": 384,
    "start_time": "2023-06-28T11:22:19.934Z"
   },
   {
    "duration": 210,
    "start_time": "2023-06-28T11:23:41.430Z"
   },
   {
    "duration": 217,
    "start_time": "2023-06-28T11:23:45.685Z"
   },
   {
    "duration": 7,
    "start_time": "2023-06-28T11:23:54.536Z"
   },
   {
    "duration": 17,
    "start_time": "2023-06-28T11:23:58.112Z"
   },
   {
    "duration": 662,
    "start_time": "2023-06-28T11:24:17.510Z"
   },
   {
    "duration": 197,
    "start_time": "2023-06-28T11:24:20.281Z"
   },
   {
    "duration": 214,
    "start_time": "2023-06-28T11:24:31.311Z"
   },
   {
    "duration": 379,
    "start_time": "2023-06-28T11:25:36.381Z"
   },
   {
    "duration": 392,
    "start_time": "2023-06-28T11:26:09.646Z"
   },
   {
    "duration": 229,
    "start_time": "2023-06-28T11:26:14.312Z"
   },
   {
    "duration": 12082,
    "start_time": "2023-06-28T11:32:30.966Z"
   },
   {
    "duration": 12864,
    "start_time": "2023-06-28T11:32:48.128Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-28T11:33:00.995Z"
   },
   {
    "duration": 17,
    "start_time": "2023-06-28T11:33:38.665Z"
   },
   {
    "duration": 19,
    "start_time": "2023-06-28T11:33:43.866Z"
   },
   {
    "duration": 11,
    "start_time": "2023-06-28T11:33:48.635Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-28T11:33:51.449Z"
   },
   {
    "duration": 10,
    "start_time": "2023-06-28T11:34:06.182Z"
   },
   {
    "duration": 352,
    "start_time": "2023-06-28T11:34:20.018Z"
   },
   {
    "duration": 11,
    "start_time": "2023-06-28T11:35:05.154Z"
   },
   {
    "duration": 1399,
    "start_time": "2023-06-28T11:35:15.345Z"
   },
   {
    "duration": 17,
    "start_time": "2023-06-28T11:35:28.862Z"
   },
   {
    "duration": 1866,
    "start_time": "2023-06-28T11:35:45.130Z"
   },
   {
    "duration": 12,
    "start_time": "2023-06-28T11:35:48.508Z"
   },
   {
    "duration": 192,
    "start_time": "2023-06-28T11:36:16.240Z"
   },
   {
    "duration": 300,
    "start_time": "2023-06-28T11:36:28.107Z"
   },
   {
    "duration": 10547,
    "start_time": "2023-06-28T11:36:43.638Z"
   },
   {
    "duration": 1458,
    "start_time": "2023-06-28T11:36:54.187Z"
   },
   {
    "duration": 17,
    "start_time": "2023-06-28T11:37:11.153Z"
   },
   {
    "duration": 1353,
    "start_time": "2023-06-28T11:37:32.762Z"
   },
   {
    "duration": 13,
    "start_time": "2023-06-28T11:37:37.307Z"
   },
   {
    "duration": 467,
    "start_time": "2023-06-28T11:38:25.412Z"
   },
   {
    "duration": 397,
    "start_time": "2023-06-28T11:38:32.405Z"
   },
   {
    "duration": 2233,
    "start_time": "2023-06-28T11:38:41.443Z"
   },
   {
    "duration": 2353,
    "start_time": "2023-06-28T11:38:51.601Z"
   },
   {
    "duration": 6,
    "start_time": "2023-06-28T11:44:08.460Z"
   },
   {
    "duration": 14943,
    "start_time": "2023-06-28T11:45:30.683Z"
   },
   {
    "duration": 5,
    "start_time": "2023-06-28T11:45:48.229Z"
   },
   {
    "duration": 331,
    "start_time": "2023-06-28T11:45:50.152Z"
   },
   {
    "duration": 208,
    "start_time": "2023-06-28T11:47:28.967Z"
   },
   {
    "duration": 388,
    "start_time": "2023-06-28T11:47:32.809Z"
   },
   {
    "duration": 263,
    "start_time": "2023-06-28T11:47:33.420Z"
   },
   {
    "duration": 14920,
    "start_time": "2023-06-28T11:47:47.418Z"
   },
   {
    "duration": 7,
    "start_time": "2023-06-28T11:48:07.015Z"
   },
   {
    "duration": 372,
    "start_time": "2023-06-28T11:48:09.819Z"
   },
   {
    "duration": 411,
    "start_time": "2023-06-28T11:48:11.763Z"
   },
   {
    "duration": 285,
    "start_time": "2023-06-28T11:48:14.118Z"
   },
   {
    "duration": 117,
    "start_time": "2023-06-28T11:48:32.175Z"
   },
   {
    "duration": 372,
    "start_time": "2023-06-28T11:48:41.366Z"
   },
   {
    "duration": 176,
    "start_time": "2023-06-28T11:50:40.024Z"
   },
   {
    "duration": 2489,
    "start_time": "2023-06-28T11:51:04.655Z"
   },
   {
    "duration": 61,
    "start_time": "2023-06-28T11:54:23.078Z"
   },
   {
    "duration": 2004,
    "start_time": "2023-06-28T11:54:28.110Z"
   },
   {
    "duration": 19611,
    "start_time": "2023-06-28T11:54:30.116Z"
   },
   {
    "duration": 15873,
    "start_time": "2023-06-28T11:54:49.730Z"
   },
   {
    "duration": 6,
    "start_time": "2023-06-28T11:55:16.419Z"
   },
   {
    "duration": 387,
    "start_time": "2023-06-28T11:55:20.511Z"
   },
   {
    "duration": 439,
    "start_time": "2023-06-28T11:55:23.721Z"
   },
   {
    "duration": 2694,
    "start_time": "2023-07-06T02:53:15.165Z"
   },
   {
    "duration": 23445,
    "start_time": "2023-07-06T02:53:17.861Z"
   },
   {
    "duration": 3424,
    "start_time": "2023-07-06T02:53:41.308Z"
   },
   {
    "duration": 1514,
    "start_time": "2023-07-06T02:53:44.734Z"
   },
   {
    "duration": 10575,
    "start_time": "2023-07-06T02:53:46.249Z"
   },
   {
    "duration": 2144,
    "start_time": "2023-07-06T02:53:56.826Z"
   },
   {
    "duration": 3216,
    "start_time": "2023-07-06T02:53:58.972Z"
   },
   {
    "duration": 1059,
    "start_time": "2023-07-06T02:54:02.190Z"
   },
   {
    "duration": 180,
    "start_time": "2023-07-06T02:54:03.250Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-06T02:54:03.432Z"
   },
   {
    "duration": 139,
    "start_time": "2023-07-06T02:54:03.442Z"
   },
   {
    "duration": 178,
    "start_time": "2023-07-06T02:54:03.583Z"
   },
   {
    "duration": 134,
    "start_time": "2023-07-06T02:54:03.762Z"
   },
   {
    "duration": 12233,
    "start_time": "2023-07-06T02:54:03.898Z"
   },
   {
    "duration": 3191,
    "start_time": "2023-07-06T02:54:16.134Z"
   },
   {
    "duration": 12287,
    "start_time": "2023-07-06T02:54:19.327Z"
   },
   {
    "duration": 3291,
    "start_time": "2023-07-06T02:54:31.615Z"
   },
   {
    "duration": 10568,
    "start_time": "2023-07-06T02:54:34.908Z"
   },
   {
    "duration": 4,
    "start_time": "2023-07-06T02:54:45.478Z"
   },
   {
    "duration": 273,
    "start_time": "2023-07-06T02:54:45.484Z"
   },
   {
    "duration": 16,
    "start_time": "2023-07-06T02:54:45.759Z"
   },
   {
    "duration": 590,
    "start_time": "2023-07-06T02:54:45.776Z"
   },
   {
    "duration": 141,
    "start_time": "2023-07-06T02:54:46.368Z"
   },
   {
    "duration": 180,
    "start_time": "2023-07-06T02:54:46.511Z"
   },
   {
    "duration": 142,
    "start_time": "2023-07-06T02:54:46.693Z"
   },
   {
    "duration": 94,
    "start_time": "2023-07-06T02:54:46.838Z"
   },
   {
    "duration": 347,
    "start_time": "2023-07-06T02:54:46.935Z"
   },
   {
    "duration": 288,
    "start_time": "2023-07-06T02:54:47.285Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-06T02:54:47.575Z"
   },
   {
    "duration": 45442,
    "start_time": "2023-07-06T02:54:47.581Z"
   },
   {
    "duration": 26855,
    "start_time": "2023-07-06T02:55:33.025Z"
   },
   {
    "duration": 26925,
    "start_time": "2023-07-06T02:55:59.882Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-06T02:56:26.808Z"
   },
   {
    "duration": 10,
    "start_time": "2023-07-06T02:56:26.818Z"
   },
   {
    "duration": 23,
    "start_time": "2023-07-06T02:56:26.829Z"
   },
   {
    "duration": 11,
    "start_time": "2023-07-06T02:56:26.853Z"
   },
   {
    "duration": 7503,
    "start_time": "2023-07-06T02:56:26.865Z"
   },
   {
    "duration": 327,
    "start_time": "2023-07-06T02:56:34.369Z"
   },
   {
    "duration": 35,
    "start_time": "2023-07-06T02:56:34.698Z"
   },
   {
    "duration": 22,
    "start_time": "2023-07-06T02:56:34.734Z"
   },
   {
    "duration": 129,
    "start_time": "2023-07-06T02:56:34.758Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-06T02:56:34.888Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-06T02:56:34.889Z"
   },
   {
    "duration": 2095,
    "start_time": "2023-07-07T03:55:27.070Z"
   },
   {
    "duration": 15910,
    "start_time": "2023-07-07T03:55:29.168Z"
   },
   {
    "duration": 3,
    "start_time": "2023-07-07T03:55:45.079Z"
   },
   {
    "duration": 1834,
    "start_time": "2023-07-07T03:55:45.084Z"
   },
   {
    "duration": 627,
    "start_time": "2023-07-07T03:55:46.932Z"
   },
   {
    "duration": 13287,
    "start_time": "2023-07-07T03:55:47.561Z"
   },
   {
    "duration": 2184,
    "start_time": "2023-07-07T03:56:00.850Z"
   },
   {
    "duration": 3295,
    "start_time": "2023-07-07T03:56:03.042Z"
   },
   {
    "duration": 954,
    "start_time": "2023-07-07T03:56:06.342Z"
   },
   {
    "duration": 242,
    "start_time": "2023-07-07T03:56:07.299Z"
   },
   {
    "duration": 16,
    "start_time": "2023-07-07T03:56:07.542Z"
   },
   {
    "duration": 174,
    "start_time": "2023-07-07T03:56:07.559Z"
   },
   {
    "duration": 240,
    "start_time": "2023-07-07T03:56:07.735Z"
   },
   {
    "duration": 178,
    "start_time": "2023-07-07T03:56:07.977Z"
   },
   {
    "duration": 12770,
    "start_time": "2023-07-07T03:56:08.157Z"
   },
   {
    "duration": 3587,
    "start_time": "2023-07-07T03:56:20.934Z"
   },
   {
    "duration": 13120,
    "start_time": "2023-07-07T03:56:24.522Z"
   },
   {
    "duration": 3797,
    "start_time": "2023-07-07T03:56:37.647Z"
   },
   {
    "duration": 12807,
    "start_time": "2023-07-07T03:56:41.446Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T03:56:54.255Z"
   },
   {
    "duration": 259,
    "start_time": "2023-07-07T03:56:54.262Z"
   },
   {
    "duration": 31,
    "start_time": "2023-07-07T03:56:54.522Z"
   },
   {
    "duration": 546,
    "start_time": "2023-07-07T03:56:54.555Z"
   },
   {
    "duration": 154,
    "start_time": "2023-07-07T03:56:55.103Z"
   },
   {
    "duration": 187,
    "start_time": "2023-07-07T03:56:55.258Z"
   },
   {
    "duration": 13,
    "start_time": "2023-07-07T03:56:55.448Z"
   },
   {
    "duration": 167,
    "start_time": "2023-07-07T03:56:55.463Z"
   },
   {
    "duration": 153,
    "start_time": "2023-07-07T03:56:55.638Z"
   },
   {
    "duration": 157,
    "start_time": "2023-07-07T03:56:55.792Z"
   },
   {
    "duration": 95,
    "start_time": "2023-07-07T03:56:55.951Z"
   },
   {
    "duration": 279,
    "start_time": "2023-07-07T03:56:56.048Z"
   },
   {
    "duration": 231,
    "start_time": "2023-07-07T03:56:56.329Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T03:56:56.562Z"
   },
   {
    "duration": 20,
    "start_time": "2023-07-07T03:56:56.568Z"
   },
   {
    "duration": 57503,
    "start_time": "2023-07-07T03:56:56.589Z"
   },
   {
    "duration": 1775,
    "start_time": "2023-07-07T04:12:50.827Z"
   },
   {
    "duration": 10831,
    "start_time": "2023-07-07T04:12:52.604Z"
   },
   {
    "duration": 3,
    "start_time": "2023-07-07T04:13:03.437Z"
   },
   {
    "duration": 1921,
    "start_time": "2023-07-07T04:13:03.442Z"
   },
   {
    "duration": 695,
    "start_time": "2023-07-07T04:13:05.365Z"
   },
   {
    "duration": 10022,
    "start_time": "2023-07-07T04:13:06.062Z"
   },
   {
    "duration": 2356,
    "start_time": "2023-07-07T04:13:16.085Z"
   },
   {
    "duration": 3224,
    "start_time": "2023-07-07T04:13:18.443Z"
   },
   {
    "duration": 1064,
    "start_time": "2023-07-07T04:13:21.670Z"
   },
   {
    "duration": 237,
    "start_time": "2023-07-07T04:13:22.736Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:13:22.975Z"
   },
   {
    "duration": 150,
    "start_time": "2023-07-07T04:13:22.985Z"
   },
   {
    "duration": 242,
    "start_time": "2023-07-07T04:13:23.137Z"
   },
   {
    "duration": 154,
    "start_time": "2023-07-07T04:13:23.380Z"
   },
   {
    "duration": 12256,
    "start_time": "2023-07-07T04:13:23.536Z"
   },
   {
    "duration": 2925,
    "start_time": "2023-07-07T04:13:35.794Z"
   },
   {
    "duration": 11317,
    "start_time": "2023-07-07T04:13:38.721Z"
   },
   {
    "duration": 3008,
    "start_time": "2023-07-07T04:13:50.040Z"
   },
   {
    "duration": 11829,
    "start_time": "2023-07-07T04:13:53.049Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T04:14:04.880Z"
   },
   {
    "duration": 251,
    "start_time": "2023-07-07T04:14:04.887Z"
   },
   {
    "duration": 17,
    "start_time": "2023-07-07T04:14:05.140Z"
   },
   {
    "duration": 597,
    "start_time": "2023-07-07T04:14:05.159Z"
   },
   {
    "duration": 149,
    "start_time": "2023-07-07T04:14:05.758Z"
   },
   {
    "duration": 136,
    "start_time": "2023-07-07T04:14:05.909Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:14:06.047Z"
   },
   {
    "duration": 185,
    "start_time": "2023-07-07T04:14:06.058Z"
   },
   {
    "duration": 187,
    "start_time": "2023-07-07T04:14:06.248Z"
   },
   {
    "duration": 139,
    "start_time": "2023-07-07T04:14:06.437Z"
   },
   {
    "duration": 82,
    "start_time": "2023-07-07T04:14:06.577Z"
   },
   {
    "duration": 331,
    "start_time": "2023-07-07T04:14:06.660Z"
   },
   {
    "duration": 266,
    "start_time": "2023-07-07T04:14:06.993Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:14:07.261Z"
   },
   {
    "duration": 7,
    "start_time": "2023-07-07T04:14:07.274Z"
   },
   {
    "duration": 44243,
    "start_time": "2023-07-07T04:14:07.282Z"
   },
   {
    "duration": 18667,
    "start_time": "2023-07-07T04:14:51.526Z"
   },
   {
    "duration": 30398,
    "start_time": "2023-07-07T04:15:10.194Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:15:40.593Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:15:40.600Z"
   },
   {
    "duration": 10,
    "start_time": "2023-07-07T04:15:40.611Z"
   },
   {
    "duration": 17,
    "start_time": "2023-07-07T04:15:40.622Z"
   },
   {
    "duration": 4,
    "start_time": "2023-07-07T04:15:40.641Z"
   },
   {
    "duration": 8476,
    "start_time": "2023-07-07T04:15:40.647Z"
   },
   {
    "duration": 363,
    "start_time": "2023-07-07T04:15:49.125Z"
   },
   {
    "duration": 23,
    "start_time": "2023-07-07T04:15:49.490Z"
   },
   {
    "duration": 22,
    "start_time": "2023-07-07T04:15:49.515Z"
   },
   {
    "duration": 147,
    "start_time": "2023-07-07T04:15:49.539Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T04:15:49.687Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T04:15:49.688Z"
   },
   {
    "duration": 1908,
    "start_time": "2023-07-07T04:40:01.746Z"
   },
   {
    "duration": 27211,
    "start_time": "2023-07-07T04:40:03.656Z"
   },
   {
    "duration": 3,
    "start_time": "2023-07-07T04:40:30.868Z"
   },
   {
    "duration": 1992,
    "start_time": "2023-07-07T04:40:30.873Z"
   },
   {
    "duration": 709,
    "start_time": "2023-07-07T04:40:32.867Z"
   },
   {
    "duration": 9922,
    "start_time": "2023-07-07T04:40:33.577Z"
   },
   {
    "duration": 2122,
    "start_time": "2023-07-07T04:40:43.500Z"
   },
   {
    "duration": 3239,
    "start_time": "2023-07-07T04:40:45.624Z"
   },
   {
    "duration": 1046,
    "start_time": "2023-07-07T04:40:48.865Z"
   },
   {
    "duration": 236,
    "start_time": "2023-07-07T04:40:49.913Z"
   },
   {
    "duration": 8,
    "start_time": "2023-07-07T04:40:50.151Z"
   },
   {
    "duration": 157,
    "start_time": "2023-07-07T04:40:50.160Z"
   },
   {
    "duration": 234,
    "start_time": "2023-07-07T04:40:50.318Z"
   },
   {
    "duration": 136,
    "start_time": "2023-07-07T04:40:50.553Z"
   },
   {
    "duration": 12165,
    "start_time": "2023-07-07T04:40:50.691Z"
   },
   {
    "duration": 2824,
    "start_time": "2023-07-07T04:41:02.857Z"
   },
   {
    "duration": 10683,
    "start_time": "2023-07-07T04:41:05.682Z"
   },
   {
    "duration": 2953,
    "start_time": "2023-07-07T04:41:16.367Z"
   },
   {
    "duration": 11996,
    "start_time": "2023-07-07T04:41:19.322Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T04:41:31.320Z"
   },
   {
    "duration": 413,
    "start_time": "2023-07-07T04:41:31.328Z"
   },
   {
    "duration": 18,
    "start_time": "2023-07-07T04:41:31.743Z"
   },
   {
    "duration": 535,
    "start_time": "2023-07-07T04:41:31.775Z"
   },
   {
    "duration": 153,
    "start_time": "2023-07-07T04:41:32.313Z"
   },
   {
    "duration": 139,
    "start_time": "2023-07-07T04:41:32.467Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:41:32.608Z"
   },
   {
    "duration": 157,
    "start_time": "2023-07-07T04:41:32.619Z"
   },
   {
    "duration": 137,
    "start_time": "2023-07-07T04:41:32.780Z"
   },
   {
    "duration": 124,
    "start_time": "2023-07-07T04:41:32.919Z"
   },
   {
    "duration": 83,
    "start_time": "2023-07-07T04:41:33.045Z"
   },
   {
    "duration": 339,
    "start_time": "2023-07-07T04:41:33.131Z"
   },
   {
    "duration": 277,
    "start_time": "2023-07-07T04:41:33.472Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T04:41:33.751Z"
   },
   {
    "duration": 17,
    "start_time": "2023-07-07T04:41:33.759Z"
   },
   {
    "duration": 43256,
    "start_time": "2023-07-07T04:41:33.777Z"
   },
   {
    "duration": 17179,
    "start_time": "2023-07-07T04:42:17.035Z"
   },
   {
    "duration": 28683,
    "start_time": "2023-07-07T04:42:34.216Z"
   },
   {
    "duration": 4,
    "start_time": "2023-07-07T04:43:02.901Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:43:02.907Z"
   },
   {
    "duration": 12,
    "start_time": "2023-07-07T04:43:02.917Z"
   },
   {
    "duration": 19,
    "start_time": "2023-07-07T04:43:02.930Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:43:02.951Z"
   },
   {
    "duration": 8193,
    "start_time": "2023-07-07T04:43:02.962Z"
   },
   {
    "duration": 301,
    "start_time": "2023-07-07T04:43:11.156Z"
   },
   {
    "duration": 30,
    "start_time": "2023-07-07T04:43:11.459Z"
   },
   {
    "duration": 23,
    "start_time": "2023-07-07T04:43:11.491Z"
   },
   {
    "duration": 156,
    "start_time": "2023-07-07T04:43:11.516Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T04:43:11.674Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T04:43:11.675Z"
   },
   {
    "duration": 8857,
    "start_time": "2023-07-07T04:48:33.843Z"
   },
   {
    "duration": 1347,
    "start_time": "2023-07-07T04:48:42.702Z"
   },
   {
    "duration": 19890,
    "start_time": "2023-07-07T04:48:44.051Z"
   },
   {
    "duration": 3,
    "start_time": "2023-07-07T04:49:03.943Z"
   },
   {
    "duration": 1890,
    "start_time": "2023-07-07T04:49:03.947Z"
   },
   {
    "duration": 730,
    "start_time": "2023-07-07T04:49:05.839Z"
   },
   {
    "duration": 9773,
    "start_time": "2023-07-07T04:49:06.570Z"
   },
   {
    "duration": 2190,
    "start_time": "2023-07-07T04:49:16.346Z"
   },
   {
    "duration": 3352,
    "start_time": "2023-07-07T04:49:18.538Z"
   },
   {
    "duration": 1058,
    "start_time": "2023-07-07T04:49:21.894Z"
   },
   {
    "duration": 237,
    "start_time": "2023-07-07T04:49:22.954Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:49:23.193Z"
   },
   {
    "duration": 142,
    "start_time": "2023-07-07T04:49:23.204Z"
   },
   {
    "duration": 233,
    "start_time": "2023-07-07T04:49:23.347Z"
   },
   {
    "duration": 145,
    "start_time": "2023-07-07T04:49:23.582Z"
   },
   {
    "duration": 11840,
    "start_time": "2023-07-07T04:49:23.728Z"
   },
   {
    "duration": 2791,
    "start_time": "2023-07-07T04:49:35.570Z"
   },
   {
    "duration": 10843,
    "start_time": "2023-07-07T04:49:38.363Z"
   },
   {
    "duration": 2937,
    "start_time": "2023-07-07T04:49:49.210Z"
   },
   {
    "duration": 11394,
    "start_time": "2023-07-07T04:49:52.149Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:50:03.545Z"
   },
   {
    "duration": 241,
    "start_time": "2023-07-07T04:50:03.551Z"
   },
   {
    "duration": 17,
    "start_time": "2023-07-07T04:50:03.793Z"
   },
   {
    "duration": 583,
    "start_time": "2023-07-07T04:50:03.812Z"
   },
   {
    "duration": 136,
    "start_time": "2023-07-07T04:50:04.397Z"
   },
   {
    "duration": 136,
    "start_time": "2023-07-07T04:50:04.534Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:50:04.674Z"
   },
   {
    "duration": 150,
    "start_time": "2023-07-07T04:50:04.686Z"
   },
   {
    "duration": 150,
    "start_time": "2023-07-07T04:50:04.838Z"
   },
   {
    "duration": 126,
    "start_time": "2023-07-07T04:50:04.989Z"
   },
   {
    "duration": 78,
    "start_time": "2023-07-07T04:50:05.117Z"
   },
   {
    "duration": 326,
    "start_time": "2023-07-07T04:50:05.197Z"
   },
   {
    "duration": 254,
    "start_time": "2023-07-07T04:50:05.525Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:50:05.781Z"
   },
   {
    "duration": 21,
    "start_time": "2023-07-07T04:50:05.787Z"
   },
   {
    "duration": 47199,
    "start_time": "2023-07-07T04:50:05.809Z"
   },
   {
    "duration": 19299,
    "start_time": "2023-07-07T04:50:53.010Z"
   },
   {
    "duration": 26312,
    "start_time": "2023-07-07T04:51:12.311Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:51:38.625Z"
   },
   {
    "duration": 23,
    "start_time": "2023-07-07T04:51:38.631Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T04:51:38.655Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T04:51:38.665Z"
   },
   {
    "duration": 7,
    "start_time": "2023-07-07T04:51:38.672Z"
   },
   {
    "duration": 8239,
    "start_time": "2023-07-07T04:51:38.681Z"
   },
   {
    "duration": 283,
    "start_time": "2023-07-07T04:51:46.922Z"
   },
   {
    "duration": 24,
    "start_time": "2023-07-07T04:51:47.207Z"
   },
   {
    "duration": 23,
    "start_time": "2023-07-07T04:51:47.233Z"
   },
   {
    "duration": 292,
    "start_time": "2023-07-07T04:51:47.257Z"
   },
   {
    "duration": 85,
    "start_time": "2023-07-07T04:51:47.551Z"
   },
   {
    "duration": 218,
    "start_time": "2023-07-07T04:51:47.638Z"
   },
   {
    "duration": 6459,
    "start_time": "2023-07-07T04:54:03.066Z"
   },
   {
    "duration": 1220,
    "start_time": "2023-07-07T04:54:09.528Z"
   },
   {
    "duration": 31029,
    "start_time": "2023-07-07T04:54:10.750Z"
   },
   {
    "duration": 2,
    "start_time": "2023-07-07T04:54:41.781Z"
   },
   {
    "duration": 1865,
    "start_time": "2023-07-07T04:54:41.785Z"
   },
   {
    "duration": 658,
    "start_time": "2023-07-07T04:54:43.652Z"
   },
   {
    "duration": 9679,
    "start_time": "2023-07-07T04:54:44.312Z"
   },
   {
    "duration": 2238,
    "start_time": "2023-07-07T04:54:53.993Z"
   },
   {
    "duration": 3410,
    "start_time": "2023-07-07T04:54:56.232Z"
   },
   {
    "duration": 1532,
    "start_time": "2023-07-07T04:54:59.644Z"
   },
   {
    "duration": 236,
    "start_time": "2023-07-07T04:55:01.178Z"
   },
   {
    "duration": 13,
    "start_time": "2023-07-07T04:55:01.416Z"
   },
   {
    "duration": 213,
    "start_time": "2023-07-07T04:55:01.431Z"
   },
   {
    "duration": 236,
    "start_time": "2023-07-07T04:55:01.646Z"
   },
   {
    "duration": 157,
    "start_time": "2023-07-07T04:55:01.884Z"
   },
   {
    "duration": 12133,
    "start_time": "2023-07-07T04:55:02.043Z"
   },
   {
    "duration": 2850,
    "start_time": "2023-07-07T04:55:14.177Z"
   },
   {
    "duration": 11055,
    "start_time": "2023-07-07T04:55:17.029Z"
   },
   {
    "duration": 3193,
    "start_time": "2023-07-07T04:55:28.085Z"
   },
   {
    "duration": 13217,
    "start_time": "2023-07-07T04:55:31.281Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:55:44.500Z"
   },
   {
    "duration": 242,
    "start_time": "2023-07-07T04:55:44.506Z"
   },
   {
    "duration": 18,
    "start_time": "2023-07-07T04:55:44.749Z"
   },
   {
    "duration": 546,
    "start_time": "2023-07-07T04:55:44.776Z"
   },
   {
    "duration": 145,
    "start_time": "2023-07-07T04:55:45.324Z"
   },
   {
    "duration": 144,
    "start_time": "2023-07-07T04:55:45.471Z"
   },
   {
    "duration": 8,
    "start_time": "2023-07-07T04:55:45.617Z"
   },
   {
    "duration": 186,
    "start_time": "2023-07-07T04:55:45.629Z"
   },
   {
    "duration": 143,
    "start_time": "2023-07-07T04:55:45.817Z"
   },
   {
    "duration": 132,
    "start_time": "2023-07-07T04:55:45.962Z"
   },
   {
    "duration": 85,
    "start_time": "2023-07-07T04:55:46.096Z"
   },
   {
    "duration": 370,
    "start_time": "2023-07-07T04:55:46.182Z"
   },
   {
    "duration": 272,
    "start_time": "2023-07-07T04:55:46.553Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T04:55:46.827Z"
   },
   {
    "duration": 13,
    "start_time": "2023-07-07T04:55:46.834Z"
   },
   {
    "duration": 41763,
    "start_time": "2023-07-07T04:55:46.849Z"
   },
   {
    "duration": 726625,
    "start_time": "2023-07-07T04:56:28.614Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.240Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.242Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.243Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.244Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.245Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.245Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.246Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.247Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.249Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.250Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.251Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.252Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T05:08:35.274Z"
   },
   {
    "duration": 9747,
    "start_time": "2023-07-07T12:36:48.890Z"
   },
   {
    "duration": 1683,
    "start_time": "2023-07-07T12:36:58.640Z"
   },
   {
    "duration": 8155,
    "start_time": "2023-07-07T12:37:00.325Z"
   },
   {
    "duration": 3,
    "start_time": "2023-07-07T12:37:08.482Z"
   },
   {
    "duration": 1821,
    "start_time": "2023-07-07T12:37:08.487Z"
   },
   {
    "duration": 741,
    "start_time": "2023-07-07T12:37:10.310Z"
   },
   {
    "duration": 10595,
    "start_time": "2023-07-07T12:37:11.056Z"
   },
   {
    "duration": 2336,
    "start_time": "2023-07-07T12:37:21.653Z"
   },
   {
    "duration": 3485,
    "start_time": "2023-07-07T12:37:23.991Z"
   },
   {
    "duration": 1128,
    "start_time": "2023-07-07T12:37:27.480Z"
   },
   {
    "duration": 1964,
    "start_time": "2023-07-07T12:37:28.610Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.578Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.580Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.581Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.582Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.584Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.585Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.586Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.587Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.588Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.589Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.590Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.592Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.593Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.594Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.595Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.596Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.597Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.599Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.599Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.600Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.601Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.603Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.604Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.605Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.606Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.607Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.608Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.609Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.610Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.611Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.612Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.613Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.655Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.656Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.657Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.658Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.661Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.667Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.671Z"
   },
   {
    "duration": 0,
    "start_time": "2023-07-07T12:37:30.673Z"
   },
   {
    "duration": 257,
    "start_time": "2023-07-07T12:38:39.787Z"
   },
   {
    "duration": 393,
    "start_time": "2023-07-07T12:38:51.088Z"
   },
   {
    "duration": 1422,
    "start_time": "2023-07-07T12:39:50.405Z"
   },
   {
    "duration": 1398,
    "start_time": "2023-07-07T12:40:08.980Z"
   },
   {
    "duration": 212,
    "start_time": "2023-07-07T12:41:05.029Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T12:41:06.531Z"
   },
   {
    "duration": 166,
    "start_time": "2023-07-07T12:41:08.748Z"
   },
   {
    "duration": 208,
    "start_time": "2023-07-07T12:41:12.732Z"
   },
   {
    "duration": 183,
    "start_time": "2023-07-07T12:41:14.403Z"
   },
   {
    "duration": 428,
    "start_time": "2023-07-07T12:43:00.583Z"
   },
   {
    "duration": 1304,
    "start_time": "2023-07-07T12:45:21.604Z"
   },
   {
    "duration": 257,
    "start_time": "2023-07-07T12:55:23.226Z"
   },
   {
    "duration": 229,
    "start_time": "2023-07-07T12:55:25.849Z"
   },
   {
    "duration": 418,
    "start_time": "2023-07-07T12:55:40.574Z"
   },
   {
    "duration": 415,
    "start_time": "2023-07-07T12:57:55.508Z"
   },
   {
    "duration": 1419,
    "start_time": "2023-07-07T12:58:19.899Z"
   },
   {
    "duration": 2833,
    "start_time": "2023-07-07T12:58:36.147Z"
   },
   {
    "duration": 10,
    "start_time": "2023-07-07T12:58:38.983Z"
   },
   {
    "duration": 7,
    "start_time": "2023-07-07T12:58:38.995Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T12:58:39.005Z"
   },
   {
    "duration": 1966,
    "start_time": "2023-07-07T12:58:39.012Z"
   },
   {
    "duration": 867,
    "start_time": "2023-07-07T12:58:40.980Z"
   },
   {
    "duration": 11247,
    "start_time": "2023-07-07T12:58:41.848Z"
   },
   {
    "duration": 2310,
    "start_time": "2023-07-07T12:58:53.097Z"
   },
   {
    "duration": 3254,
    "start_time": "2023-07-07T12:58:55.409Z"
   },
   {
    "duration": 1065,
    "start_time": "2023-07-07T12:58:58.667Z"
   },
   {
    "duration": 1402,
    "start_time": "2023-07-07T12:58:59.735Z"
   },
   {
    "duration": 450,
    "start_time": "2023-07-07T12:59:01.139Z"
   },
   {
    "duration": 13250,
    "start_time": "2023-07-07T12:59:01.591Z"
   },
   {
    "duration": 3434,
    "start_time": "2023-07-07T12:59:14.843Z"
   },
   {
    "duration": 11963,
    "start_time": "2023-07-07T12:59:18.279Z"
   },
   {
    "duration": 3668,
    "start_time": "2023-07-07T12:59:30.244Z"
   },
   {
    "duration": 13251,
    "start_time": "2023-07-07T12:59:33.914Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T12:59:47.168Z"
   },
   {
    "duration": 262,
    "start_time": "2023-07-07T12:59:47.179Z"
   },
   {
    "duration": 32,
    "start_time": "2023-07-07T12:59:47.443Z"
   },
   {
    "duration": 594,
    "start_time": "2023-07-07T12:59:47.478Z"
   },
   {
    "duration": 175,
    "start_time": "2023-07-07T12:59:48.074Z"
   },
   {
    "duration": 199,
    "start_time": "2023-07-07T12:59:48.253Z"
   },
   {
    "duration": 18,
    "start_time": "2023-07-07T12:59:48.454Z"
   },
   {
    "duration": 161,
    "start_time": "2023-07-07T12:59:48.474Z"
   },
   {
    "duration": 144,
    "start_time": "2023-07-07T12:59:48.637Z"
   },
   {
    "duration": 132,
    "start_time": "2023-07-07T12:59:48.783Z"
   },
   {
    "duration": 89,
    "start_time": "2023-07-07T12:59:48.919Z"
   },
   {
    "duration": 325,
    "start_time": "2023-07-07T12:59:49.011Z"
   },
   {
    "duration": 273,
    "start_time": "2023-07-07T12:59:49.338Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T12:59:49.613Z"
   },
   {
    "duration": 15,
    "start_time": "2023-07-07T12:59:49.620Z"
   },
   {
    "duration": 58465,
    "start_time": "2023-07-07T12:59:49.637Z"
   },
   {
    "duration": 626374,
    "start_time": "2023-07-07T13:00:48.104Z"
   },
   {
    "duration": 25787,
    "start_time": "2023-07-07T13:11:14.480Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T13:11:40.269Z"
   },
   {
    "duration": 11,
    "start_time": "2023-07-07T13:11:40.280Z"
   },
   {
    "duration": 19,
    "start_time": "2023-07-07T13:11:40.293Z"
   },
   {
    "duration": 7,
    "start_time": "2023-07-07T13:11:40.314Z"
   },
   {
    "duration": 23,
    "start_time": "2023-07-07T13:11:40.323Z"
   },
   {
    "duration": 9818,
    "start_time": "2023-07-07T13:11:40.348Z"
   },
   {
    "duration": 342,
    "start_time": "2023-07-07T13:11:50.168Z"
   },
   {
    "duration": 26,
    "start_time": "2023-07-07T13:11:50.512Z"
   },
   {
    "duration": 40,
    "start_time": "2023-07-07T13:11:50.540Z"
   },
   {
    "duration": 374,
    "start_time": "2023-07-07T13:11:50.582Z"
   },
   {
    "duration": 74,
    "start_time": "2023-07-07T13:11:50.958Z"
   },
   {
    "duration": 265,
    "start_time": "2023-07-07T13:11:51.033Z"
   },
   {
    "duration": 6560,
    "start_time": "2023-07-07T16:14:20.439Z"
   },
   {
    "duration": 1499,
    "start_time": "2023-07-07T16:14:27.001Z"
   },
   {
    "duration": 11073,
    "start_time": "2023-07-07T16:14:28.502Z"
   },
   {
    "duration": 3,
    "start_time": "2023-07-07T16:14:39.577Z"
   },
   {
    "duration": 2766,
    "start_time": "2023-07-07T16:14:39.582Z"
   },
   {
    "duration": 705,
    "start_time": "2023-07-07T16:14:42.349Z"
   },
   {
    "duration": 9381,
    "start_time": "2023-07-07T16:14:43.056Z"
   },
   {
    "duration": 2098,
    "start_time": "2023-07-07T16:14:52.439Z"
   },
   {
    "duration": 3270,
    "start_time": "2023-07-07T16:14:54.539Z"
   },
   {
    "duration": 1046,
    "start_time": "2023-07-07T16:14:57.811Z"
   },
   {
    "duration": 1272,
    "start_time": "2023-07-07T16:14:58.859Z"
   },
   {
    "duration": 426,
    "start_time": "2023-07-07T16:15:00.133Z"
   },
   {
    "duration": 12281,
    "start_time": "2023-07-07T16:15:00.561Z"
   },
   {
    "duration": 3085,
    "start_time": "2023-07-07T16:15:12.843Z"
   },
   {
    "duration": 11713,
    "start_time": "2023-07-07T16:15:15.931Z"
   },
   {
    "duration": 3116,
    "start_time": "2023-07-07T16:15:27.646Z"
   },
   {
    "duration": 13985,
    "start_time": "2023-07-07T16:15:30.763Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T16:15:44.749Z"
   },
   {
    "duration": 214,
    "start_time": "2023-07-07T16:15:44.755Z"
   },
   {
    "duration": 17,
    "start_time": "2023-07-07T16:15:44.971Z"
   },
   {
    "duration": 518,
    "start_time": "2023-07-07T16:15:44.990Z"
   },
   {
    "duration": 158,
    "start_time": "2023-07-07T16:15:45.509Z"
   },
   {
    "duration": 145,
    "start_time": "2023-07-07T16:15:45.669Z"
   },
   {
    "duration": 7,
    "start_time": "2023-07-07T16:15:45.815Z"
   },
   {
    "duration": 159,
    "start_time": "2023-07-07T16:15:45.832Z"
   },
   {
    "duration": 134,
    "start_time": "2023-07-07T16:15:45.992Z"
   },
   {
    "duration": 118,
    "start_time": "2023-07-07T16:15:46.127Z"
   },
   {
    "duration": 68,
    "start_time": "2023-07-07T16:15:46.248Z"
   },
   {
    "duration": 318,
    "start_time": "2023-07-07T16:15:46.317Z"
   },
   {
    "duration": 270,
    "start_time": "2023-07-07T16:15:46.638Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-07T16:15:46.910Z"
   },
   {
    "duration": 39,
    "start_time": "2023-07-07T16:15:46.916Z"
   },
   {
    "duration": 39971,
    "start_time": "2023-07-07T16:15:46.956Z"
   },
   {
    "duration": 30416,
    "start_time": "2023-07-07T16:16:26.929Z"
   },
   {
    "duration": 23115,
    "start_time": "2023-07-07T16:16:57.347Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T16:17:20.464Z"
   },
   {
    "duration": 6,
    "start_time": "2023-07-07T16:17:20.471Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-07T16:17:20.479Z"
   },
   {
    "duration": 8,
    "start_time": "2023-07-07T16:17:20.489Z"
   },
   {
    "duration": 10,
    "start_time": "2023-07-07T16:17:20.498Z"
   },
   {
    "duration": 8314,
    "start_time": "2023-07-07T16:17:20.509Z"
   },
   {
    "duration": 320,
    "start_time": "2023-07-07T16:17:28.831Z"
   },
   {
    "duration": 24,
    "start_time": "2023-07-07T16:17:29.152Z"
   },
   {
    "duration": 22,
    "start_time": "2023-07-07T16:17:29.178Z"
   },
   {
    "duration": 316,
    "start_time": "2023-07-07T16:17:29.201Z"
   },
   {
    "duration": 70,
    "start_time": "2023-07-07T16:17:29.531Z"
   },
   {
    "duration": 224,
    "start_time": "2023-07-07T16:17:29.602Z"
   },
   {
    "duration": 88,
    "start_time": "2023-07-07T16:37:19.692Z"
   },
   {
    "duration": 16257,
    "start_time": "2023-07-08T10:56:22.517Z"
   },
   {
    "duration": 3401,
    "start_time": "2023-07-08T10:56:38.790Z"
   },
   {
    "duration": 14757,
    "start_time": "2023-07-08T10:56:42.195Z"
   },
   {
    "duration": 16,
    "start_time": "2023-07-08T10:56:56.956Z"
   },
   {
    "duration": 3981,
    "start_time": "2023-07-08T10:56:56.985Z"
   },
   {
    "duration": 1663,
    "start_time": "2023-07-08T10:57:00.972Z"
   },
   {
    "duration": 22441,
    "start_time": "2023-07-08T10:57:02.637Z"
   },
   {
    "duration": 5393,
    "start_time": "2023-07-08T10:57:25.086Z"
   },
   {
    "duration": 7738,
    "start_time": "2023-07-08T10:57:30.482Z"
   },
   {
    "duration": 2539,
    "start_time": "2023-07-08T10:57:38.223Z"
   },
   {
    "duration": 2970,
    "start_time": "2023-07-08T10:57:40.764Z"
   },
   {
    "duration": 500,
    "start_time": "2023-07-08T10:57:43.736Z"
   },
   {
    "duration": 21655,
    "start_time": "2023-07-08T10:57:44.239Z"
   },
   {
    "duration": 7491,
    "start_time": "2023-07-08T10:58:05.896Z"
   },
   {
    "duration": 20406,
    "start_time": "2023-07-08T10:58:13.392Z"
   },
   {
    "duration": 9165,
    "start_time": "2023-07-08T10:58:33.801Z"
   },
   {
    "duration": 18537,
    "start_time": "2023-07-08T10:58:42.969Z"
   },
   {
    "duration": 9,
    "start_time": "2023-07-08T10:59:01.560Z"
   },
   {
    "duration": 827,
    "start_time": "2023-07-08T10:59:01.571Z"
   },
   {
    "duration": 79,
    "start_time": "2023-07-08T10:59:02.400Z"
   },
   {
    "duration": 1675,
    "start_time": "2023-07-08T10:59:02.488Z"
   },
   {
    "duration": 404,
    "start_time": "2023-07-08T10:59:04.169Z"
   },
   {
    "duration": 490,
    "start_time": "2023-07-08T10:59:04.580Z"
   },
   {
    "duration": 16,
    "start_time": "2023-07-08T10:59:05.073Z"
   },
   {
    "duration": 507,
    "start_time": "2023-07-08T10:59:05.096Z"
   },
   {
    "duration": 487,
    "start_time": "2023-07-08T10:59:05.606Z"
   },
   {
    "duration": 401,
    "start_time": "2023-07-08T10:59:06.103Z"
   },
   {
    "duration": 246,
    "start_time": "2023-07-08T10:59:06.563Z"
   },
   {
    "duration": 998,
    "start_time": "2023-07-08T10:59:06.812Z"
   },
   {
    "duration": 401,
    "start_time": "2023-07-08T10:59:07.812Z"
   },
   {
    "duration": 8,
    "start_time": "2023-07-08T10:59:08.215Z"
   },
   {
    "duration": 62,
    "start_time": "2023-07-08T10:59:08.225Z"
   },
   {
    "duration": 137149,
    "start_time": "2023-07-08T10:59:08.289Z"
   },
   {
    "duration": 6896,
    "start_time": "2023-07-09T10:53:46.637Z"
   },
   {
    "duration": 1656,
    "start_time": "2023-07-09T10:53:53.544Z"
   },
   {
    "duration": 12162,
    "start_time": "2023-07-09T10:53:55.202Z"
   },
   {
    "duration": 4,
    "start_time": "2023-07-09T10:54:07.366Z"
   },
   {
    "duration": 2045,
    "start_time": "2023-07-09T10:54:07.371Z"
   },
   {
    "duration": 721,
    "start_time": "2023-07-09T10:54:09.418Z"
   },
   {
    "duration": 9853,
    "start_time": "2023-07-09T10:54:10.142Z"
   },
   {
    "duration": 2138,
    "start_time": "2023-07-09T10:54:19.997Z"
   },
   {
    "duration": 3357,
    "start_time": "2023-07-09T10:54:22.138Z"
   },
   {
    "duration": 1065,
    "start_time": "2023-07-09T10:54:25.498Z"
   },
   {
    "duration": 1228,
    "start_time": "2023-07-09T10:54:26.565Z"
   },
   {
    "duration": 11837,
    "start_time": "2023-07-09T10:54:27.795Z"
   },
   {
    "duration": 2853,
    "start_time": "2023-07-09T10:54:39.634Z"
   },
   {
    "duration": 11339,
    "start_time": "2023-07-09T10:54:42.489Z"
   },
   {
    "duration": 3071,
    "start_time": "2023-07-09T10:54:53.832Z"
   },
   {
    "duration": 11914,
    "start_time": "2023-07-09T10:54:56.905Z"
   },
   {
    "duration": 11,
    "start_time": "2023-07-09T10:55:08.821Z"
   },
   {
    "duration": 205,
    "start_time": "2023-07-09T10:55:08.833Z"
   },
   {
    "duration": 20,
    "start_time": "2023-07-09T10:55:09.040Z"
   },
   {
    "duration": 466,
    "start_time": "2023-07-09T10:55:09.061Z"
   },
   {
    "duration": 140,
    "start_time": "2023-07-09T10:55:09.531Z"
   },
   {
    "duration": 145,
    "start_time": "2023-07-09T10:55:09.673Z"
   },
   {
    "duration": 14,
    "start_time": "2023-07-09T10:55:09.819Z"
   },
   {
    "duration": 143,
    "start_time": "2023-07-09T10:55:09.835Z"
   },
   {
    "duration": 138,
    "start_time": "2023-07-09T10:55:09.980Z"
   },
   {
    "duration": 125,
    "start_time": "2023-07-09T10:55:10.120Z"
   },
   {
    "duration": 71,
    "start_time": "2023-07-09T10:55:10.246Z"
   },
   {
    "duration": 321,
    "start_time": "2023-07-09T10:55:10.320Z"
   },
   {
    "duration": 264,
    "start_time": "2023-07-09T10:55:10.643Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-09T10:55:10.909Z"
   },
   {
    "duration": 15,
    "start_time": "2023-07-09T10:55:10.916Z"
   },
   {
    "duration": 42688,
    "start_time": "2023-07-09T10:55:10.933Z"
   },
   {
    "duration": 160735,
    "start_time": "2023-07-09T10:55:53.623Z"
   },
   {
    "duration": 20936,
    "start_time": "2023-07-09T10:58:34.360Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-09T10:58:55.297Z"
   },
   {
    "duration": 11,
    "start_time": "2023-07-09T10:58:55.304Z"
   },
   {
    "duration": 20,
    "start_time": "2023-07-09T10:58:55.316Z"
   },
   {
    "duration": 5,
    "start_time": "2023-07-09T10:58:55.338Z"
   },
   {
    "duration": 4,
    "start_time": "2023-07-09T10:58:55.345Z"
   },
   {
    "duration": 8957,
    "start_time": "2023-07-09T10:58:55.350Z"
   },
   {
    "duration": 357,
    "start_time": "2023-07-09T10:59:04.309Z"
   },
   {
    "duration": 25,
    "start_time": "2023-07-09T10:59:04.668Z"
   },
   {
    "duration": 44,
    "start_time": "2023-07-09T10:59:04.695Z"
   },
   {
    "duration": 376,
    "start_time": "2023-07-09T10:59:04.741Z"
   },
   {
    "duration": 74,
    "start_time": "2023-07-09T10:59:05.119Z"
   },
   {
    "duration": 242,
    "start_time": "2023-07-09T10:59:05.195Z"
   }
  ],
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}