notebooks/Run pipeline.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "import urllib\n",
    "import re\n",
    "from tqdm import tqdm \n",
    "%matplotlib inline\n",
    "import sys\n",
    "import seaborn as sns\n",
    "\n",
    "sys.path.append('../code/')\n",
    "from utils import *\n",
    "from etl import *\n",
    "import warnings\n",
    "warnings.simplefilter('ignore')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Connect to the database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "database_config = load_config('../misc/database_config.yaml')[0]\n",
    "    \n",
    "engine = get_engine(database_config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run ETL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "transactions\n",
      "users\n",
      "fraudsters\n",
      "currency_details\n",
      "countries\n"
     ]
    }
   ],
   "source": [
    "!python3 -W ignore ../code/etl.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Retrieve the data from database and process it. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "df_transactions = get_transactions(engine)\n",
    "\n",
    "df_users = get_users(engine, df_transactions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data exploration"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Missing features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "currency                      0\n",
       "amount                        0\n",
       "state                         0\n",
       "created_date                  0\n",
       "merchant_category        465586\n",
       "merchant_country         205596\n",
       "entry_method                  0\n",
       "user_id                       0\n",
       "type                          0\n",
       "source                        0\n",
       "id                            0\n",
       "amount_usd                53323\n",
       "merchant_country_name    209550\n",
       "is_crypto                     0\n",
       "exponent                      0\n",
       "is_fraudster                  0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_transactions.isna().sum(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Impute missing amount_usd with an API"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "df_transactions = fill_missing_amount_usd(df_transactions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.PairGrid at 0x7f8d3376db38>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 624.125x540 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_users_no_na = df_users[~df_users['is_fraudster'].isna()].dropna()\n",
    "df_users_no_na.loc[:, 'is_fraudster'] = df_users_no_na.replace({False:'Normal',True:'Fraudster'})\n",
    "sns.pairplot(df_users_no_na[['transaction_count', 'total_amount_usd', 'is_fraudster', 'state',\n",
    "                            'median_amount_usd']], hue='is_fraudster',\n",
    "            dropna=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAx8AAAEYCAYAAAA9Pks0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3XeYnOV57/HvPTPbtUVdq4YkENWAhAQCDAZjx8bdyYkTHBMTmxxsx/g6sX0cl+TEBpvECUmIDYkxTiCAAYELIBAgRO/qBXWtCkLSCkm7K2nr1Pv8Me+KlVhJo9VO2dnf57rmmneetzz3CPaduedp5u6IiIiIiIhkWyjfAYiIiIiIyOCg5ENERERERHJCyYeIiIiIiOSEkg8REREREckJJR8iIiIiIpITSj5ERERERCQnlHyIiIiIiEhOKPkQCZjZa30453NmttbMns9CPJPMbNVR9teZ2V/1d73HiOlyM9tvZsuDxzPHOP4vzOy2XMUnIqJ7+THj+WiPe3ibma0Ptu/JVQwyuCn5EAm4+8V9OO1a4K/c/YM9C80s0j9RHVUdcFwfWJZ2on/3L7v7tODx4RO8lohIv9K9/OjcfV73PRxYDHwheP3Fw+rIxXuXQUjJh0jAzNqC53ozeyn4JWiVmV16hOP/HrgEuN3Mbg5+5f+NmT0GPG1mQ8zsWTNbamZvmtlngvMO+RXMzP6vmf0o2J5hZivM7HXg6z2OOcvMFgYxrTSzqcBPgZODspuD475jZouCY27oUd9aM/tPYCkwIQv/dp8yswVmtszMnjGz0b0c87ng33OFmb0UlIWDf7vumL/S37GJyOCie/kJ/dv9pZnNNrPHgSfN7MNm9kiP/beb2dXB9vlm9qKZLTGzJ3u774v0RlmtyHv9GTDP3W8yszBQ2dtB7n6jmV0B/F93X2xmfwFcBJzj7s3Br0Z/6O4HzGwE8IaZzTlG3XcB33D3F7s/hAJfBX7m7veZWSkQBr4HvC/49Qoz+wgwFbgAMGCOmX0A2AacBnzJ3d/z65qZ3QJ88PByYLa7/7SX8kvNbHmw/Rt3vwl4BbjQ3d3M/hL4G+Dbh53398BH3X2HmdUFZdcC+939fDMrA141s6fdfcvR/pFERDKge3nake7lR3IRMM3dW8ys19bt4H79M+DT7r7XzL4A/Bi47jjqkUFKyYfIey0C7jSzEuARd19+rBN6mO/uzcG2Af8QfGikgHHAEX8ZMrNaoM7dXwyK7gU+Fmy/DvytmY0Hfu/uG83s8Et8JHgsC14PIf0Btg14y93f6K1ed//mcbw/SHe7+uRhZeOBB82sHigFekseXgX+x8weAn7fI+ZzzOyPg9e1QcxKPkTkROle3jdPu3vLMY45AzgLeCaIPwxs76f6pcgp+RA5jLu/FHzIfAK418xudvdMB+K199j+AjASmOHucTPbCpQDCQ7t8lgePBvgR4jpfjNbEMQ0L2hd2HzYYQb8o7v/8pBCs0mHxcVh+/vj17JbgX9z9zlmdjnwo17ew1fNbFbwHpab2bQg5m+4+7wM6xERyYju5Qcdb8tHzzqO9h5XunuvXdlEjkbJh8hhzOwkYIe7/8rMqoDzgL7MAlIL7A4+rD4InBSUvwOMMrPhQBvwSeApd99n6ZmkLnH3V0h/4HXHNAXY7O4/D7bPAVYA1T3qmwf82Mzuc/c2MxsHxI8VZD/9WlYL7Ai2r+ntADM72d0XAAvM7FOk+yvPA75mZs8F/06nkv63P+IHrIhIJnQv7xdvAWcFXcSqgCuAZ4A1wDgzu8DdFwb7p7r76izEIEVGyYfIe10OfMfM4qQ/UL549MOP6D7gMTNbDCwH1gEEH2A3AgtIdy9a1+OcL5HuJtBB+gOo258CVwcx7QJuDPoiv2rpAY9Puvt3zOwM4PWgGbwNuBpI9jH+4/Ej4DdmtgN4A5jcyzE3B4MrDXiW9AfuSmASsNTSQe8BPpuDeEWk+F2O7uUnxN23BAPO3wQ2kB7ojrtHg+6yPzezatLfJ/8VUPIhx2TuvbYMioiIiIiI9CtNtSsiIiIiIjmhblciGQgGCJYdVvzn7v5mPuIREZHjp3u5SP6p25WIiIiIiOSEWj56ceWVV/pTTz2V7zBERAab9yx40Be6h4uI5EVG93CN+ejF3r178x2CiIj0ke7hIiKFS8mHiIiIiIjkhJIPERERERHJCSUfIiIiIiKSE0o+REREREQkJ5R8iIiIiIhITij5EBERERGRnFDyISIiIiIiOaHkQ0REREREckLJh4iIiIiI5ISSD5EiMnv2bFatWpXvMERERER6Fcl3ACLSf26//XZqamqYM2dOvkMREREReQ+1fIgUmQMHDuQ7BBEREZFeKfkQEREREZGcUPIhIiIiIiI5oeRDRERERERyQsmHSJFw93yHICIiInJUSj5EikQymcx3CCIiIiJHpeRDpEjE4/F8hyAiIiJyVEo+RIqEkg8REREpdEo+RIpELBbLdwgiIiIiR6XkQ6RIKPkQERGRQqfkQ6RIKPkQERGRQqfkQ6RIaMyHiMjA1dDQwPXXf4OVK1fmOxSRrFLyIVIklHyIiAxcK1asYNWqN3nhhRfyHYpIVin5ECkSiUQi3yGIiEgfdXV1AZBKpfIciUh2KfkQKRJaZFBEZODq7OwEdC+X4qfkQ0RERCTPOjo6gHeTEJFipeRDREREJM/a2toAaG1tzXMkItmV9+TDzLaa2ZtmttzMFgdlw8xsvpltDJ6HBuVmZj83swYzW2lm5/W4zjXB8RvN7Joe5TOC6zcE51ru36VI9ul/bRGRgevAgQOHPIsUq7wnH4EPuvs0d58ZvP4e8Ky7TwWeDV4DfAyYGjyuA34B6WQF+CEwC7gA+GF3whIcc12P867M/tsRyb1QqFD+nEVE5Hjt378fgH3Bs0ixKtRvK58B7g627wY+26P8Hk97A6gzs3rgo8B8d2929xZgPnBlsK/G3V93dwfu6XEtkaKi5ENEZOBqam4GoKVlX54jEcmuQvi24sDTZrbEzK4Lyka7eyNA8DwqKB8HvN3j3O1B2dHKt/dS/h5mdp2ZLTazxXv27DnBtySSe0o+ZDDTPVwGMnenpaUFgK7OjoPT7ooUo0L4tvJ+dz+PdJeqr5vZB45ybG+d2r0P5e8tdL/D3We6+8yRI0ceK2aRgqPkQwYz3cNlIGtrayMei5GsHAbA3r178xyRSPbk/duKu+8MnncDD5Mes/FO0GWK4Hl3cPh2YEKP08cDO49RPr6XcpGiowHnIiID0+7d6a85yapRh7wWKUZ5TT7MrMrMqru3gY8Aq4A5QPeMVdcAjwbbc4AvBrNeXQjsD7plzQM+YmZDg4HmHwHmBftazezCYJarL/a4lkhRUfIhIjIwNTY2ApCsGQPArl278hmOSFZF8lz/aODh4EtTBLjf3Z8ys0XAQ2Z2LbAN+Fxw/BPAx4EGoAP4EoC7N5vZj4FFwXE3untzsP014H+ACuDJ4CFSdJR8iIgMTAeTj+oxYHbwtUgxymvy4e6bgXN7KW8CPtRLuQNfP8K17gTu7KV8MfC+Ew5WREREJAt27NiBlZThkQoor2b79u3HPklkgMr7mA8RERGRweztt98mWVoDZiRKa3hr27Z8hySSNUo+RIqEul2JiAxMm7dsIVleC0CqvJa3336bVCqV56hEskPJh4iIiEietLa20tLcTKqiDoBUxVDisZjGfUjRUvIhUiTU8iEiMvBs2bIFgGTF0OC57pBykWKj5EOkSKTnYxARkYFk8+bNQLrFo+fzpk2b8haTSDYp+RApEko+REQGnoaGhvRMV6VV6YJwCVTU0tDQkN/ARLJEyYdIkVDyISIy8GzYuJFE+TDo0XU2Xj6UDRs35jEqkexR8iFSJDQziojIwJJIJNiyeTPJymGHlKcqh/POrl20trbmKTKR7FHyIVIklHyIiAws27ZtIx6Pk6wacUh5smo4gLpeSVFS8iFSJJR8iIgMLOvXrwfSLR09db/esGFDzmMSyTYlHyJFQsmHiMjAsmHDBixcQqq85pByL6nAyoco+ZCipORDpEgo+RARGVjWrV9PonIY2Hu/jsXKh7F23bo8RCWSXUo+RIqEZrsSERk4EokEDQ0NJCtH9Lo/VTWCnTt20N7enuPIRLJLyYdIkVDyISIycGzbto14LEbysPEe3ZIa9yFFSsmHSJFQtysRkYHj4GDzqiO3fICSDyk+Sj5EioSSDxGRgePdwea1ve73kgqsbMjBJEWkWCj5ECkSSj5ERAaOtevWkagcfsjK5oeLVQxn7TolH1JclHyIFAklHyIiA0MikWBTQ8MRx3t0S1UNp3HnDtra2nIUmUj2KfkQKRKJRCLfIYiISAa2bt3a68rmh0tq3IcUISUfIkVCyYeIyMCwLli/I9PkY+3atVmPSSRXlHyIFAklHyIiA8O6deuwkjK8rOboB0bKoaLmYLIiUgyUfIgUiWQyeXBb4z9ERArXqtVriFeOOOpg827xihGsXrMmB1GJ5IaSD5Ei0TP56LktIiKFo6Ojg7e2biFZNTKj45NDRtLc1MTu3buzHJlIbij5ECkSPRMOdcESESlM69atw91JDhmV0fHJqvRxa9T6IUVCyYdIkVC3KxGRwrdq1SqAjFs+UpXDsXCEN998M5thieSMkg+RItEz4VC3KxGRwrRy5Uq8chhEyjI7IRQiUTmSFStXZjcwkRwpiOTDzMJmtszMHg9eTzazBWa20cweNLPSoLwseN0Q7J/U4xrfD8rXm9lHe5RfGZQ1mNn3cv3eRHKlZ1crtXyIiBSeRCLBm6tWEa/KrMvVwfOqR7OpoYH29vYsRSaSOwWRfAD/B+g5ifU/Abe4+1SgBbg2KL8WaHH3U4BbguMwszOBq4CzgCuB/wwSmjDwH8DHgDOBzwfHihQdjfkQESlsGzZsINrVRbKm/rjOS1aPwd3V9UqKQt6TDzMbD3wC+K/gtQFXAL8NDrkb+Gyw/ZngNcH+DwXHfwaY7e5Rd98CNAAXBI8Gd9/s7jFgdnCsSNGJxWIHt+PxeB4jERGR3ixbtgxIJxPHIzlkFITCLF26NBthieRU3pMP4N+BvwG6+4kMB/a5e/dPt9uBccH2OOBtgGD//uD4g+WHnXOkcpGi0zP5iEajeYxERER6s3jJErxyGF5ScXwnhiIkq0axeMmS7AQmkkN5TT7M7JPAbnfv+dfU24o7fox9x1veWyzXmdliM1u8Z8+eo0QtUph6JhxKPmSw0T1cCl00GmXVm28Srz6+LlfdEjX1bN60iX379vVzZCK5le+Wj/cDnzazraS7RF1BuiWkzswiwTHjgZ3B9nZgAkCwvxZo7ll+2DlHKn8Pd7/D3We6+8yRIzOb/k6kkHR1dfW6LTIY6B4uhe7NN98kHo+TqBnbp/O7z1ui1g8Z4PKafLj79919vLtPIj1g/Dl3/wLwPPDHwWHXAI8G23OC1wT7n3N3D8qvCmbDmgxMBRYCi4CpwexZpUEdc3Lw1kRyTsmHiEjhWrx4MYRCxz3eo1uqagQWKUtfR2QAixz7kLz4LjDbzH4CLAP+Oyj/b+BeM2sg3eJxFYC7rzazh4A1QAL4ursnAczsemAeEAbudPfVOX0nIjnS1dWFYxiu5ENEpMAsWLiQZNVoCJf07QIWIlZdz8KFi3B30vPtiAw8BZN8uPsLwAvB9mbSM1UdfkwX8LkjnH8TcFMv5U8AT/RjqCIFqaurK71oVaJLyYeISAFpbm5my+bNJMbPOKHrJGvH0bT1Vd566y0mTZrUP8GJ5Fi+x3yISD/pikZJRcoBDTgXESkk3V2lEjUnNuFm9/nqeiUDmZIPkSIRjUbxSOnBbRERKQxLlizBSipIVQ4/oet42RCoqGXRokX9FJlI7in5ECkSsVgMwunkQyuci4gUBndnwcJFxKrHQD+M04hV17N8+QotJisDlpIPkSIRjyfwUCTY1oeSiEgh2LZtG/tamkn2cYrdwyWrxxKNdrF27dp+uZ5Irin5ECkSyWQSQuF3t0VEJO+WLVsGQKKPiwseLlGTvs7y5cv75XoiuabkQ6RIJJNJsBBgSj5ERArE8uXLsbIheFl1/1wwUoZXDWeZkg8ZoJR8iBQJT6VwDMxIr70pIiL55O4sX7GSWNWofhnv0S1eNZrVq1ZrfJ8MSEo+RIpEMpVKf7iZkUql8h2OiMig19jYmB7vUT26X6+brB5FLBaloaGhX68rkgtHXWTQzM472n53X9q/4YjIiTLU8iEiUghWrVoFQHLIqH69bnLI6IPXP/300/v12iLZdqwVzv81eC4HZgIrAAPOARYAl2QvNBHpE0PJh4hIAVi7di0WLiFVMbRfr+ulVVhZlWa8kgHpqN2u3P2D7v5B4C3gPHef6e4zgOmA2vpECoh19yd2CIXUo1JEJN/WrFlLonJ4MBlI/4pVjmD16jX9fl2RbMv0r+F0d3+z+4W7rwKmZSckEemLkBm446jVQ0Qk3+LxOJs2bSJZNSIr109VjWDXrkZaW1uzcn2RbMk0+VhrZv9lZpeb2WVm9itAbX0iBSQUDmM4eIpI5Fg9KkVEJJu2bt1KIhEnWZmd5KP7uuvXr8/K9UWyJdPk40vAauD/AH8NrAnKRKRAhMNhSCXBPb0tIiJ5s2HDBgCSVcOzcv3u627cuDEr1xfJlox+HnX3LuAW4BYzGwaMD8pEpEBEIhGIJt7dFhGRvNm4cSMWKcXLarJTQaQcK69W8iEDTkYtH2b2gpnVBInHcuAuM/u37IYmIsejpKQES8UBJR8iIvm2YeNGEhVD+3VxwcPFyoeyPmhhERkoMu12VevuB4A/Au4KZrz6cPbCEpHjFYlEsFTy4LaIiORHKpViU8MmkhXZ6XJ1sJ7K4ezcsYOOjo6s1iPSnzJNPiJmVg/8CfB4FuMRkT4qiUQgqW5XIiL5tnPnTqLRLlKVw7JaT6pyGO7Oli1bslqPSH/KNPm4EZgHNLj7IjObAqiToUgBCUfC4Gr5EBHJt4aG9FJoySwnH8nK4YfUJzIQZDrg/DfAb3q83gz8r2wFJSLHL2QhLFjZ3LLYx1hERI6uoaEBLNTvK5sfzkursJIyJR8yoGSUfJjZXfDelcvc/cv9HpGI9Ek639ACgyIi+bZx40a8og5CWZ723IxE+TDNeCUDSqZ9M3qO8ygH/hDY2f/hiEhfpdwPzqririRERCRfNmxsIFGe3VaPbsnKoWza3EAymdQaTzIgZNrt6nc9X5vZA8AzWYlIRPoklUzilv7gSSaTeY5GRGRw2r9/Py3NTSTHT8lJfcmKYcTfidHY2Mj48eNzUqfIich0wPnhpgIT+zMQETkx8UQCQunfE5R8iIjkx+bNmwFIVeam5aN7Rq1NmzblpD6RE5XpmI9WDu1Mvgv4blYiEpE+icXieJB8xOPxPEcjIjI4vfXWWwBZH2zeLVVee0i9IoUu025X1Ufbb2Znufvq/glJRPoiFo/h4VJAyYeISL5s3boVi5TiJZW5qTBcgpVXs3Xr1tzUJ3KC+trt6nD39tN1RKSPYtEYhEsAiEajeY5GRGRw2rFjB8mymoMTgORCvLSGt7dvz1l9Iieiv5KPPv2FmVm5mS00sxVmttrMbgjKJ5vZAjPbaGYPmllpUF4WvG4I9k/qca3vB+XrzeyjPcqvDMoazOx7J/Y2RQpXLBbDQxEsXEIsFst3OCIig9Lb23eQLD1qh5F+lyqrZscOTUIqA0N/JR99ndczClzh7ucC04ArzexC4J+AW9x9KtACXBscfy3Q4u6nALcEx2FmZwJXAWcBVwL/aWZhMwsD/wF8DDgT+HxwrEjRicWi6TnlQ2G1fIiI5IG707R3L15aldt6S6voaG+jq6srp/WK9EV/JR994mltwcuS4OHAFcBvg/K7gc8G258JXhPs/5Cll3L+DDDb3aPuvgVoAC4IHg3uvtndY8Ds4FiRohOPxfBQGEIRtXyIiORBW1sb8XiMVGmOxnsEuuvbu3dvTusV6Yv+Sj76/E0naKFYDuwG5gObgH3unggO2Q6MC7bHAW8DBPv3A8N7lh92zpHKe4vjOjNbbGaL9+zZ09e3I5IXiUSCVCqVnmo3FFLLhww6uodLIWhtbQXAI2U5rbe7vra2tmMcKZJ/GSUfZvbs0crc/cK+BuDuSXefBown3VJxRm+HdVd7hH3HW95bHHe4+0x3nzly5MhjBy5SQA7ObmUh3MKa7UoGHd3DpRC0t7enN0Ilua04mOlQyYcMBEedatfMyoFKYISZDeXdL/M1wNj+DMTd95nZC8CFQJ2ZRYLWjfFA9yiq7cAEYLuZRYBaoLlHebee5xypXKRoJBLphkK3MG4hdbsSEcmD7h9+PBTOab1u6d+Suz8LRArZsVo+vgIsAU4Pnrsfj5IeyH1CzGykmdUF2xXAh4G1wPPAHweHXRPUBzAneE2w/zl396D8qmA2rMmkV2BfCCwCpgazZ5WSHpQ+50TjFik0B1s6QiHcQlrhXEQkD9JfSfqmbNsblG17o49n2wnXL5IrR235cPefAT8zs2+4+61ZqL8euDuYlSoEPOTuj5vZGmC2mf0EWAb8d3D8fwP3mlkD6RaPq4I4V5vZQ8AaIAF83d2TAGZ2PTAPCAN3ajFEKUYHf+2yEE5Iv36JiORBONzd4nH8SUCoo7nvFXvqsPpFClemK5zfamYXA5N6nuPu95xI5e6+EpjeS/lm0uM/Di/vAj53hGvdBNzUS/kTwBMnEqdIoetu6XAMN1PLh4hIHlRUVABgydz+AGSpdH3l5eU5rVekLzJKPszsXuBkYDnQ/a3GgRNKPkSkf6RS6V+9MAMzEko+RERy7t3kI7fj7rrrq6zM7RS/In2RUfIBzATOdHUmFClI7/5pGmB4dzIiIiI5U1dXB4AlcrvYX3d9Q4cOzWm9In2R6Tofq4Ax2QxERPruYPJh6eRDRERyr7S0lIrKSizemdN6Ld6JmVFbW5vTekX6ItOWjxHAGjNbCBxcvczdP52VqETkhKiNUkQkP0aMGMGBAx05rdNiHdTU1BKJZPq1TiR/Mv2/9EfZDEJE+pep8UNEJC9GjxrFW01bc1pnKN6OFteUgSLT2a5ezHYgIiIiIgPdqFGjCL+5Nqd1RuKdjB49Kad1ivRVprNdtfLupNWlQAnQ7u412QpMRDJnauoQESkII0eOxKPtkEpBKNOhtSfG1PIhA0imLR/VPV+b2WfpZR0OEcmP1GGzW6U06ENEJC+GDRsGpGeg8tIcTH2bSuDxKMOHD89+XSL9oE8pubs/AlzRz7GISB+9m3yk1/lIaZ0PEZG8OJh85GjGK4trml0ZWDLtdvVHPV6GSK/7oZ9WRQpEIhGspmsh3ELE4rldXVdERNKGDBkC5G6hwe56uusVKXSZznb1qR7bCWAr8Jl+j0ZE+qSrK/3Ll4cjEIoQjeZ2gSsREUnrXuWcVDw3FaYSh9YrUuAyHfPxpWwHIiJ919kZNO+HIng48u5rERHJqVAwyNxyNPbOPN3tNhwO56Q+kROV0ZgPMxtvZg+b2W4ze8fMfmdm47MdnIhkprW1FQCPlOHhMtrb295d9VxERHKmewye52gWQrf0V7mkxvrJAJHpgPO7gDnAWGAc8FhQJiIF4GDyES6DSBmJeFytHyIiedDW1pbeCJfmpsJwCQDt7e25qU/kBGWafIx097vcPRE8/gfQhNIiBaK5uRkALyknVZLu97tv3758hiQiMii1tLQA4JHynNTXXU93vSKFLtPkY6+ZXW1m4eBxNdCUzcBEJHPNzc1YaUV6tqsg+Whq0p+oiEiuNTY2ApAqrcpJfR4px8IRdu7cmZP6RE5UpsnHl4E/AXYBjcAfB2UiUgCamppIlaQXs1LyISKSP5s3b4bymoPdobLOjGT50HS9IgNAprNdbQM+neVYRKSP9uzZQzJoevcgCenuiiUiIrnh7qxavZp4xbCc1puoHMa6detJJBJEIpmuoiCSH5kuMjgZ+AYwqec57q6ERKQANDW34CXp1W09Ug5mSj5ERHJsx44d7N2zh+RJp+S03mRNPZ171rN+/XrOOuusnNYtcrwyTY8fAf6b9CxXqeyFIyLHy93Zv38fqRH16QIzrLRCyYeISI699tprACRqxuW03kTNWDDjtddeU/IhBS/TMR9d7v5zd3/e3V/sfmQ1MhHJSGdnJ8lE4pCZVTxSfnD6XRERyY2n588nVTUSL6/JbcWRchI1Y5k//xmt8SQFL9Pk42dm9kMzu8jMzut+ZDUyEcnIwSQj8u6c8slQiZIPEZEc2rRpEw0bNxIbNiUv9ceHnczu3e+wfPnyvNQvkqlMu12dDfw5cAXvdrvy4LWI5FH3wlLeY0ErD5XS2qYFp0REcuW3v/0tFi4hPiK34z26JYZNwrYv4je//S3Tp0/PSwwimcg0+fhDYIq7x7IZjIgcv+6VzD307p+zhyN0dCj5EBHJhb179zJ//jNEh50MkbL8BBGK0DXiVF577TW2bdvGxIkT8xOHyDFk2u1qBVCXzUBEpG+6k4+ec8p7uOTdchERyar777+fRDJBbMz78hpHfPSZWCjM3Xffndc4RI4m0+RjNLDOzOaZ2ZzuRzYDE5HMHGz56LmgVaiErs6uPEUkIjJ47Nq1i0fnzCE2fGruB5ofxksq6Bp5Bs8+95wWHZSClWm3qx9mNQoR6bO2tjYAPHRoy0dXVyfJZJJwOJyv0EREit4dd9xBKgWxsdPyHQoAsfpzKG/awG3/8R/867/8C2aW75BEDpFRy0fP6XX7c6pdM5tgZs+b2VozW21m/ycoH2Zm881sY/A8NCg3M/u5mTWY2cqeM26Z2TXB8RvN7Joe5TPM7M3gnJ+b/gqlyBxMPnr0M/Zwert7MLqIiPS/VatW8dxzz9E1+iy8bEi+w0mLlNE5ZhpLlyzh9ddfz3c0Iu+RUfJhZhea2SIzazOzmJklzexAP9SfAL7t7mcAFwJfN7Mzge8Bz7r7VODZ4DXAx4CpweM64BdBfMNIt87MAi4AftidsATHXNfjvCv7IW6RgrF//34wg56zXQWJyP79+/MVlohIUUskEvzrv/0blA0hVn9OvsM5RHzU6XhFHf/+s5/T1aUuuFJYMh3zcRvweWAjUAH8ZVB2Qty90d3DzNztAAAgAElEQVSXBtutwFpgHPAZoHu01N3AZ4PtzwD3eNobQJ2Z1QMfBea7e7O7twDzgSuDfTXu/rqnV925p8e1RIpCc3MzVlqZTkACXlJxcJ+IiPS/hx9+mC2bN9M5/oJDJvwoCKEwnRMvYvc7u/j1r3+d72hEDpFp8oG7NwBhd0+6+13A5f0ZiJlNAqYDC4DR7t4Y1NsIjAoOGwe83eO07UHZ0cq391LeW/3XmdliM1u8Z8+eE307IjnT1NREqsfq5gBeWgmkp38UGQx0D5dcamxs5Fe/+i8SteNJDD0p3+H0KllTT3z4Kdx//wNs2rQp3+GIHJRp8tFhZqXAcjP7ZzP7JlDVX0GY2RDgd8Bfu/vRunP1Nl7D+1D+3kL3O9x9prvPHDly5LFCFikYu3a9Q6Lk0D/HVPBaX8JksNA9XHLF3bn55n8hnkzRddLFh7Q6F5quiRfg4VJ++k//RCKRyHc4IkDmycefB8deD7QDE4D/1R8BmFkJ6cTjPnf/fVD8TtBliuB5d1C+Pai723hg5zHKx/dSLlI09uzZg5ce9ltApBSLlCr5EBHpZ3PnzmXp0iV0jp9ZOIPMjyRSTsfEWWzcsIEHH3ww39GIAJnPdvWWu3e5+wF3v8HdvxV0wwLAzH7Xl8qDmaf+G1jr7v/WY9ccoHvGqmuAR3uUfzGY9epCYH/QLWse8BEzGxoMNP8IMC/Y1xoMmDfgiz2uJTLgdXR00NHRfrCbVU9eWqnkQ0SkH+3atYtbb7st3aVp5On5DicjiWFTiA+dxJ133aW1P6QgZDzm4xim9PG895NuVbnCzJYHj48DPwX+wMw2An8QvAZ4AtgMNAC/Av4KwN2bgR8Di4LHjUEZwNeA/wrO2QQ82cdYRQpOU1MT8G43q54SkQolHyIi/SSVSvGPP/0psXiSzkmXFHR3q8NFT7qIVKiEm276B+LxeL7DkUEu00UGj6XXcRTHPMn9FXoflwHwoV6Od+DrR7jWncCdvZQvBt7Xl/hECl1LSwvw7uxWPXmkkubmllyHJCJSlH7/+9+zYvlyuiZdgpdV5zuc4+IlFXRMvIhNDc9xzz33cO211+Y7JBnE+qvlQ0TyoHsdDz9stisALyln3759uQ5JRKTobN26lV/+8pck6iYQHzE13+H0SWLoJOLDT+G+++5j9erV+Q5HBrH+Sj4GTtujSBFpbW0FDl3dvJuHS4lGuzTDiYjICYjH49z44x+TIEzXpPcPqO5Wh+uaeCGp0ip+/JOf0NHRke9wZJDqr+Tju/10HRE5Dm1tbQB4LwtceaT0kGNEROT43XnnnWzetImOk96Pl7x3co8BJVJKx6RL2dXYyG23nfBa0SJ9ctQxH2b2Jr2P5zDSQzDOIb3xdBZiE5FjOPjLVW/JRyhd1tnZSV1dXS7DEhEpCsuWLeOB2bOJjTy1YBcTPF7J6jFEx5zDE088wUUXXcSll16a75BkkDnWgPNP5iQKEemTrq4uCIXBemnEDBKSrq6uHEclIjLwHThwgJ/85CYoryE6YVa+w+lXsXHTKWndyU//6Z857bTTGDVqVL5DkkHkqN2ugvU9jvjIVZAi0ruuri6sl1YPAA+lf1uIRqO5DElEZMBzd/75n/+ZppZm2idf1mvr8oAWCtMx5TI6Oru46aabSCaT+Y5IBpGMxnwEi/QtMrM2M4uZWdLMDmQ7OBE5umg0mm756E2QfKjlQ0Tk+MyZM4dXXnmFrrEzSFWNyHc4WeHltXRMmMWKFSu4//778x2ODCKZDji/Dfg8sBGoAP4SuDVbQYlIZqLR6MEWjsN5kJSo5UNEJHObNm3i1ltvI1k7jviY4l4mLDFiKvFhU7jzzjtZuXJlvsORQSLj2a7cvQEIu3vS3e8CPpi9sEQkE0dLPlDyISJyXDo6Ovjhj35EMlRC5+QPDOhpdTNiRteki/HyGm644UatDSU5kWny0WFmpcByM/tnM/smUJXFuEQkA9FoFLfeu125ul2JiGTM3bn55n9h+/bttE++DC+pyHdIuREupX3yZTS3tPATjf+QHMg0+fjz4NjrgXZgAvBH2QpKRDLT0dFB6ohjPjTblYhIph5++GGef/45omPPI1lTn+9wcipVNYLOCbNYvGgR99xzT77DkSKXafLxWXfvcvcD7n6Du38LTcMrknft7R0H1/M4XPfCg1rFVkTk6FasWMFtt/0HibqJxOrPyXc4eREfeRrx4adw99138+qrr+Y7HClimSYf1/RS9hf9GIeI9EFbextEynrfGYoARnt7e05jEhEZSN555x3+3//7e1Jl1YNjnMeRBOM/UlUj+fFPfsJbb2lFBcmOoyYfZvZ5M3sMmGxmc3o8XgCachKhiBxRW2sbHi7tfacZVlLGgQOaFVtEpDft7e1897vfo7Wjk/aTr4DIEe6ng0UoQsfJHySaNL7zN9/VAHTJimOtcP4a0AiMAP61R3kroDnZRPKoq6uLaLQLLyk/4jFeUq4PDxGRXiQSCX50ww1sfWsrHVM/QqqiLt8hFQQvG0LbyVfA+qf4wQ/+lltu+TfKyo7Qwi7SB5mscP6Cu18ErAOqg8d2d0/kIkAR6V1LSwsAqciRZ2RJhstpamrOVUgiIgNCKpXi5ptvZtHChXRNvIhk7bh8h1RQUkNG0TH5UtasWc0NN9xAIqGvfNJ/Ml3h/HPAQuBzwJ8AC8zsj7MZmIgc3a5duwDw0iPPep0qrWLXO7tyFZKISMFzd37xi18wb948omOnEx91er5DKkiJYZPpmnghr732GjfffDOpVCrfIUmROFa3q25/B5zv7rsBzGwk8Azw22wFJiJH1518pMqqj3hMqnQIze9sIRaLUVo6yPsyi8ig5+7ccccd/OY3vyE26kxiY6flO6SCFh99JpaIMm/ePCKRCN/+9rcJhTJen1qkV5n+HxTqTjwCTcdxrohkwebNm7FwBC8bcsRjUhVDSaVSbNu2LYeRiYgUnu4WjwceeIDYyNOJTpw1eGe2Og6xsdOI1p/L3Llz1QIi/SLTlo8nzWwe8EDw+k+BJ7ITkohkYsOGjSTL68CO/DtAqnJYcOwGTjnllFyFJiJSUBKJBLfccgtz584lNuoMohMvVOKRKTNi484DM5588kk6Ojr4wQ9+oEHo0meZtl448EvgHOBc4I6sRSQix9TV1cWq1atIDBlz1ONS5bVYaQXLli3LUWQiIoWlo6ODH/zt3zJ37lyi9ecWZOJRtu0Nwh1NhDuaqFj3BGXb3sh3SIcKEpCuCefz4osv8q1vf1vTuEufZZp8/IG7/97dv+Xu33T3h4GPZTMwETmy5cuXk0wkSNSOPfqBZsSGjGHBwkWarUREBp2dO3dy/Te+wcKFC+k66WJi42cUXOIBEOpoxpJxLBkn0rqLUEdhzlIYH3M2nVMuZ/WatXz1q19jy5Yt+Q5JBqBjLTL4NTN7EzjNzFb2eGxB63yI5M3TTz+NlZSTrK4/5rGJoZM5sH8fS5YsyUFkIiKFYcGCBfzv/30dW97aTscpH9asVv0kMXwKHad+lMa9LXz1q1/jueeey3dIMsAcq+XjfuBTwJzgufsxw92vznJsItKLAwcO8PLLLxMdOhlC4WMen6ibgJWUM3fu3BxEJyKSX7FYjF/+8pd893vfo40yWs/4FMm6CfkOq6gkq8fQesan6Syt4cYbb+SWW26hs7Mz32HJAHGsRQb3u/tWd/98sOBg96Mw2wNFBoFHHnmEeDxOfGSGv+KFwnQNn8rLL7/Mzp07sxuciEgebdq0ieu+8tX0jFbDp9J22ifw8pp8h1WUvLSS9lM/Rmz0WTz66By+fO1fsnr16nyHJQOApssVGUC6urr4zW9/R6J2AqnKoRmfFx99Fm7Ggw8+mMXoRETyIxqNcvfdd3Pdddfx1o5ddEz9MNHJl0A400k9pU9CYaITZ9Fx2pU0Nu3n+uuv5xe/+AUdHR35jkwKWN6TDzO708x2m9mqHmXDzGy+mW0MnocG5WZmPzezhmDsyXk9zrkmOH6jmV3To3yGmb0ZnPNzswIcaSaSoccee4zWA/uJ1Z99XOd5aSWx4afw+Ny57NmzJ0vRiYjk3uuvv84Xr7mGu+66i67aibSe+VmSdRPzHdagkqypp/WszxIdPpUHH3yQq//8z3n++edx93yHJgUo78kH8D/AlYeVfQ941t2nAs8GryE9w9bU4HEd8AtIJyvAD4FZwAXAD7sTluCY63qcd3hdIgNCNBrlvvvvJ1lTT7L66FPs9iZWfy7JZJLZs2dnIToRkdzasmUL3//+9/n+97/POweidJx2JV0nfxAvKc93aINTuJTo5EtoP+OTNHXBDTfcwDe/9S3Wr1+f78ikwOQ9+XD3l4DDx5B8Brg72L4b+GyP8ns87Q2gzszqgY8C89292d1bgPnAlcG+Gnd/3dPp9z09riUyoDz99NPsa2khWj+tT+d7WTXxYScz57HH2LdvXz9HJyKSGzt37uSmm27iy1/+Mm8sWkJ0/Exaz/g0yZpjTD0uOZEaMoq2Mz5F10kXsWLVWr7yla/w93//97z11lv5Dk0KRKF2hhzt7o0A7t5oZqOC8nHA2z2O2x6UHa18ey/lIgNKKpXi/gdmk6oa0adWj26x+rMpaWrg0Ucf5Zprrjn2CSIiBaKxsZHZs2fz2OOP425ER7+PaP3ZEFFLR8GxEPFRZxAffjKlu1bx8quv8/LLL/ORj3yEP/uzP+Okk07Kd4SSR4WafBxJb+M1vA/l772w2XWku2cxcaL6ikphWbp0KY07dxCdctkJLZCVqhhKonYcc+Y8xtVXX004fOypekUGAt3Di9e6det48MEHeeGFF3AzYsNPJTZ2Gl5ame/Q5FjCpcTGnUd81JmUNq7k6fnPMm/ePC6++GKuuuoqzj77bDQUd/Ap1OTjHTOrD1o96oHdQfl2oOdk3eOBnUH55YeVvxCUj+/l+Pdw9zuAOwBmzpypEVJSUObOnYuVlJMYOumIx5RtewOA6MQLj3qt+IjTaNr0HIsXL2bWrFn9GaZI3ugeXlwSiQQLFy5k9uwHWblyBRYppWv0+4iPPhMvrcp3eHKcvKSc6MQL0q3vu9fy+qKlvPbaa5x2+ulc9ad/yiWXXEJJSUm+w5QcKdTkYw5wDfDT4PnRHuXXm9ls0oPL9wcJyjzgH3oMMv8I8H13bzazVjO7EFgAfBG4NZdvROREdXR08PLLrxAddspRFxUMdWS2/E73ooNPP/20kg8RKSi7d+9m7ty5PP74XJqa9kLZELomXEB85KkQLs13eHKCvKSC2LjziI05h5KmjazfupobbriBmto6PvHxj/GJT3yC8ePHH/tCMqDlPfkwswdIt1qMMLPtpGet+inwkJldC2wDPhcc/gTwcaAB6AC+BBAkGT8GFgXH3dhjIcSvkZ5RqwJ4MniIDBivvfYaiUScxLDJ/XPBUJho3UReefVVotEoZWVl/XNdEZE+SCQSvPHGGzz2+OMsXLAAdydRO574KVeQqJ0IobzPjSP9LRxJjwkZeRrh/TtI7FnPA7Nn88ADDzD9vPP49Kc+xcUXX6zPpyKV9+TD3T9/hF0f6uVYB75+hOvcCdzZS/li4H0nEqNIPs1/5hkoqyI5ZHS/XTMxbArRPRt4/fXXufzyy/vtuiIimUilUqxatYpnnnmG555/nrbWViitJDrmHOIjT8XLqvMdouSChUjWTaCzbgIWa6dk70aWrd7AsqU3UFFRyeWXX8aHPvQhpk+frjGKRSTvyYeIHFlLSwuLFi4kOuqsExpofrhk9Rgoq2LevHlKPkQkJ9ydTZs28eyzz/L0/Gdo2rsHC5cQq51AfOoskjXj1coxiHlpFbGx04jVn0P4QCPx5s08Nf9ZnnzySerqhvKhD13Bhz/8YU4//XQNUh/glHyIFLDHHnuMVCpFYsQp/XthCxEdOoUFCxawc+dOxo7V/Pgi0v+SySRr1qzh5Zdf5qWXX2ZXYyOYkagZR3zKZSTqJkJYA42lBwuRrB1HsnYcXamLiOx7m3jTZn738CP87ne/Y/iIkVz2gUu59NJLOfvss4lE9FV2oNF/MZEC1dHRwYMP/YZE3QRSFUOPfcJxio8+i/Lda7jvvvv4zne+0+/XF5HBKRqNsmzZMl5++WVefuVVDuzfB6EQiep6EiddTGLoSXhJRb7DlIEgFCExbDKJYZPpSkSJ7NvGrpa3ePiROfz+97+nakg1l7z/Yi655BLOP/98ysu15stAoORDpEDde++9tLe1Ej3jsqxc30sriY44lSeffIo//MM/5JRT+rl1RUQGjV27drFgwQIWLFjAkiVLiUa7sEgpsZpxJKZMI1E3XrNVyYmJlJEYMZXEiKl0JeNE9u8gvu8tnn72BebNm0dJSQnTpk/nwlmzmDVrlmbNKmBKPkQK0Jo1a5g9ezaxEaeSGjIqa/VEx02nbN9b/MM//iO/vP12zbMuIhmJxWKsXLmSBQsW8Pobb7D97bfTO8qridVMIlE3gWTN2KNODy7SZ+ESEsMmkRg2ia5UinDbLiIt21j05noWLVzIrbfeypj6ei668EJmzZrFtGnT1CpSQJR8iBSYlpYWbrjxRry0iujEC7JbWaScjokXsbnhWW677Tb++q//WgP5ROQ9UqkUW7ZsYcmSJSxZsoRly5YTi0UhFCY5ZDTxCReQrB1Pqry2XyfHEDmmUIhkzViSNWOJAtZ1gMj+7Wzfv4NH5jzGww8/TCRSwjnnnMPMmTOYMWMGp5xyimbPyiMlHyIFpLOzk+997/vs3rOX9lOvzEk3hcTQk4iNeR+PPvoo9fX1XHXVVVmvU0QKX2NjI0uXLmXJkiUsWryE1gP70zsq6ojVTSFROz49c54GjEsB8fIa4uVnEh99Jp2pBOHWd4js386SdVtYunQJAFVVQ5gx4zxmzEgnI+PGjdMPbzmk5EOkQHR0dPB3f/d3rN+wns6Tr8hqd6vDRcefj0Xbuf322yktLeWP/uiPcla3iBSGvXv3snz5cpYvX87iJUvSM1MBVlZFbMgYEpPPIVkzFi+tynOkIhkKRQ7OnBUFLN6Rnsb3wE5eWrCUl156CYARI0cyc8YMpk2bxrRp0xgzZkx+4y5ySj5ECkBzczN/893v0tDQQOekS0gMPSm3AZjRNeVSbFOSn//85zQ3N3PttdfqlyCRItbU1HQw2ViydCk7d+wAwCJlxIeMIjFxFsmasaTK69SVSoqCl1SSGH4yieEn0+WORQ8QOdBI44GdzHv2BZ566ikARo0ew4zzpjNt2jSmT5/OqFG5+zFwMFDyIZJnq1ev5oYbbmRPUxMdp3yYZN2E/AQSitB5yhWUbX2NX//612zfvp1vf/vbVFdrpWGRYrB3715WrFjB8uXLWbpsGTu2bwd6JBsTLiBZPYZU5TAwLfYnRc4ML68lXl5LfNTpdLkT6mwh3LqLHQca2fPM8zz55JMAjB4zhhnnnce5557Lueeeq5aRE6TkQyRPEokE9957L/feey+p0iraT70yp12temUhopPej5dX88KLL/HmqtX83d/+gOnTp+c3LhE5brt372b58uWsWLGCpUuX0di4EwCLlBKvGk1iwvkkq+uVbIgAmJGqHEaqchjx0Wf2SEYa2X5gF7vnP8cTTzwBwMhRozlv+rRDummpp0DmlHyI5MGyZcv4z//8BRs3biA+/GS6Jl4EkQKZA9+MWP25JKrHwtaX+Oa3vsWnP/UprrnmGoYPH57v6ETkCHbv3s2yZcsOtmy8s2sXkG7ZiA0ZTVItGyKZOyQZOeuQlpGdrY3sfe4l5s2bB8DwESMPJiPTp0+nvr5eychRKPkQyaENGzZwx69+xeJFi6BsCJ0nX05i2JR8h9Wr1JCRtJ7xacq2L2HOY4/z5FNP8Sef+xxXXXWVumKJFIDuMRvLli07dIB4STmxqlEkJ8wiWTOGVMVQJRsiJ6q3lpGufYQPNLKrdRfPvPgK8+fPB2DkqFHMOO88pk+frjEjvVDyIZJlyWSSxYsX89hjj/PKKy9jJeV0TTif+KgzIFTgf4LhEqInXUhs9JmU7VjKfffdx8OPPMJnP/MZPv7xj2sFWZEcam1tZdmyZSxdupTFS5ay/e1tQM8xG93JxjANEBfJNjNSFUNJVQx9TzKys7WRvT0GsNfXj2XGjHQyMnPmTGpra/McfH6Zu+c7hoIzc+ZMX7x4cb7DkAFu+/btPPXUUzzx5JM0NzWlk44RpxEbc3ZWulhVrEv3Re08/eP9fu1uoY4mynYuI7LvbXDnrPe9j09+4hNcdtllVFZWZq1eGTT65RtzsdzDU6kUGzZsYOHChSxcuJA1a9aQSqWwcEk62ageS7JGYzaKQcXqRxiS6uCTn/wkjz/+OG2hSjrP+my+w5IT4U6os5nwgUYirY2UtL2DJ2KYGaeeehoXXjiLCy64gNNOO41IpMB/iMxcRvdwJR+9KJYPLsmtVCrFpk2bWLp0Ka+8+ipvrlwJZiRqxhEfcSqJugkQys6KqmXb3qBk70YAkpXDSVUOIzrxwqzUBWCxDkqaGihr2gid+ykrK+fyyy/j/PPPZ/r06RobIn016JOP/fv3s2DBAhYuXMiCBQtpbT0AQKpqBPGa9HoFyapREFKyUUwqVzzEn376Sq6//npuvfVWHnpsHh3n/km+w5L+5ClC7XuJ7N9ByYEdhNr3gDtVVUOYOXMGs2bNYtasWQP98zOje3jRpFoiuebu7NixgyVLlqRXAV66lLbW1vTOyqFEx80gPuKUnCzIFepoxpJxACKtu0hkuT4vrSRWfw6xMWcTbttNbO8Gnn72hYOD7yZMnMj5M2dyXjA1ocaIiBzZgQMHeOWVV3ju+edZumRJunWjtIJY9VgSU6anF/Yrqch3mJJFHinl8ccfx92ZO3cuHlFLctGxEKkho4gNGUVs3HRIRIkc2EFs/w5eemMxL774ImbGOeecwxVXXMEHPvABhg4dmu+os0ItH70YyL+aSXa4O/v27WPr1q1s2bKF9evXs3jJUpr27gHAyoakVwCuSXeDyPUKwBXrniDSuuvg60T1mKx2v+qVpwh1NBM5sJNwayMlbbvxZBwzY+qpp3Le9OlMmTKFSZMmcdJJJ1FWVpbb+GQgGDQtH62trbz66qs8//zzLFq8mFQyCeU1ROsmkRg2iVTlcI3bGEQK4h4u+RPMpBVp2UpZy1bo3IeZMW36dK744Ae59NJLqaury3eUmVDLh0hftLS0sHXr1oOPzVu2sGXLlndbNQArqUhPXXnSVBI19XhZjb4oWIhU1QhiVSOg/hw6U0nC7XsIH9jJ2p2NbNjwEHgqfagZo0aP4eQpk5k8eTKTJk1i0qRJTJw4UUmJFLXdu3fz0EMPMeexx4hFo1BeTWzkmcSHTVbCITJYBTNpxSqHERs7PZ2ING9h2dpNLFu6lH//2c/4+Mc+xuc//3nGjh2b72hPmJIPGXRSqRQtLS3s2bPn4GPbtm1s3rKFzZu30Bb0sYb0LDKJ8lqSFfWk6s4kVVFHqmJouguEviQcXShMsnoMyeoxxABSKULR/YQ69xHq3Mf2zhbeWb6W115/45CkZPSY+oNJydixYxk5cuTBR1VVbluURPrL9u3buf/++5k3bx7JVIr4sCnEppxBqmqk7iUi8q6eici48wh1NFOyZx2PzX2Cxx9/nCuuuIIvfOELTJlSmNP0Z0LJhxSVRCJBU1PTIYnFnj172Lt3L+/s3s3u3btpbm5Od3HowSKlJMvrSJSPIVV3ejB9Xh1eUqkvBv0lFDo4LWG3LoBUklD0AKHOFkKd+3i7cx+7lq3h1ddeg8O6hZaXVzB8xAjGjB51SFIyYsSIg9u1tbVa3EkKyoMPPsjtt98OFiY64lRiY96Hl2kclIgcgxmpquFEq95PbOx0St9ZxXMvvMSzzz7L1VdfzbXXXjsgP++UfMiAEI1G2b9/P/v372ffvn20tLSwd+/eg8lFOrHYw4H9+zh8HJOFI3jpEBKRCry0htSo9JiMVEkVXlqJl1bhkXIlGfkSCh8xKbF4B6FYOxZrx2IdxOLttLZ38Pb6bYRXrcNjHe9JUCIlJQwfPoLRo0YyatS7Scrw4cOpq6ujtraWuro6qqurCYezM/uYSLf77ruPX/3qV8SHTiJ60kUaOC4ifeKllUQnXEC0/lzK3l7Ir3/9a+LxOF/96lcHXAKi5ENyLplM0traejCZ6E4oDt9uaWmhZd9+Wg8cIBrt6vVaVlKGl1SRKKkgVTISrz8pnViUVuEllaRKqyBcqsRiIAqF8bJqkkf7hdhTWLwTi7UTinVg8fRzZ6ydnVveIbJxKx5th1TyPaeaGZVVVdTW1jF0aB11QVJS2+P58O2KiooBd5OX/HnuuefSicewKf+/vXuNreO87zz+fWbmzLnwXHinKFG8U2ItlbIdVTZtYx3b3UYJDF0WdjYJsCiwdYMg9XaBvmmLIu+7bxoskAbbYlugaIF1gr5os8raDnZt2Lt2vZWrWHRk60KJpCiS4k0kz43nMmeevpjhISmRkiyTPLz8P8DBzDycOXwEUnP4m+dGrvPfyFocQogvzwqSb38ODIuf/OQnNDc3c+bMzloTRsKH+FK01iwuLq4KEmuFirm5eeb8/Uw6dU/rxBJlBcAKUTJDuFYQbcXR1Q1oK4QOhLytFcS1wmg7AmZgi//FYltRhtdyZVfhrneO1ignhypmUU4eVcx5x06OvJNjPpvnZnIWyx1HOXl0YbE8BuVuViBALBanujpBbU3NfYNKPB4nHo9j2xu/oKTYGaampgDIH/wNCR5CiI2jFPmW49jTl5mcnKx0bb4wCR+iTGtNJpNhYWGBZDK5arvylUwmmZubZyGZJJVcoFS696kyAEqhAmG0FcIxbS84hPajY8sh4u5QgSG/kmKDKYUOhB++u4vWUCqWA4oXXHJecHFyLDo5pu/kUFM3MUsFlJNDF/Prvl0wFCIeT1BdnaDaDyYrAwlGVVAAABU2SURBVMpaW5nxa3d44oknAAjMXKPQfExaYIUQGyYwex205sknn6x0Vb4w+UtvFysUCkxNTa0KDXdv5+eXWiSSpNOpewZiL1Mo2wsJJcP2WyXq0Q0H1ggS3r50dxI7klJg2WjLRhN/uGtcF1XKoYr5FaHF2y84edJOjonJLMbYHKabh2IO7RTWfbtgMEQ0FqO6upqa6rWDytKrpqaG2tpaDFnxetvp7u6m79gxBi5ewErdZrGtHx1KVLpaQogdTOXThG5+hDV/k67ubvr6+ipdpS9MwscOprXmzp07jI+PMzExwcTEBOPj44yPjzM2PsHcndm1uzeVWySClMwgrhlE243ohlZ0ILg6TEiQEOLBDANtRLzZ0R6WW0KV8isCy/K24ORIO3kmpxcxJuYxXO887azdwmIFAjQ17aPlwH72799Pc3Mzzc3N5f1IRFZLrgTTNPnhn/0Z586d47/9xV9gXvoHCo2PUazvXjXBghBCPIjKLRCYvU5o8lcETJP/+L3v8corr2BZO+9P+Z1X40eglDoJ/FfABP671vpPK1ylh5bNZrl9+/Y9AePW2DiTt29TLK5+eqqCVTh2FNeuxm1uwQ3G7gkTEiR2oVKBUCjEyy+/zLlz50iX1n+qLrYJw3yEwOJ6gWVlWClkMfJphtMpbv1qEONffnlPq0o0FufAgf3sXxFIlsJJQ0PDjvzw2ilM0+T06dM899xz/PjHP+add97Bvv0pOlJDobodp7YDN7wjVi4WQmwxlUsSmBvCnhtGZWYBePqZZ/jPv//77Nu3r8K1e3S7/hNHKWUCfw78W+AWcF4p9TOt9WeVrdmDjYyM8Nprr1EsFstlyrJxgzGcQBRd24MbjHkBIxjDDUZlzMQepZwCL596mddffx2tNT/9n29XukpiMxgG2njA+BWtoZTHyKe99VPyKQr5NPPjSa6MjEPuvVUD6o8ePcqPfvSjLaj83lZXV8cPfvADvv/97/P+++/zf955h0u/+gQ9/kt0pJZCTTtOfL+3yrkhU0ALsSdpFyMzi5WaIDA3jJGZAeBwby8vvfjv+epXv0pjY2OFK/nl7YW/VE8Ag1rrGwBKqTeA08C2Dx/Nzc18+9vf5r3332dkeBgAHYhQjO3HqT5IKdooM6gIALRlc+7cObTW/PznP0db0s1mz1IKrBCuFcKtqvfKnDzWwi30/Ci2333LtCyefPJJTp86Vdn67jF1dXWcPXuWs2fPMj09zXvvvcc777zLZ59dIDh2AWVaOJEGnFgTpWiTd5+XWf12PTdSi856T7ZLkTrcSG2FayS2hOtgpqcx05OYqdsEMtPokvfAubunh9986RWef/55mpubK1zRjaXWm/J0t1BKvQKc1Fq/5h//B+AprfXrd533XeC7AK2trV8ZGRnZ8rrez8TEBB9++CEffPAhn1z8BLdUQlk2OhDGNW1v3IYVRC9tV75WlEmXq90pfPl/YaVul4+d2D4We79RwRqJTeOWUP4sW8rJg+Pvl/J+Vyz/VcpjlooYpTw6nwaticUTPPfsMzzzzDN85Stf2Y5jQR755rTd7+EPMjc3x6effsrAwACfXLzI9cFBb8yeUriROpyoH0aq6tF2ldzHd6HgzY8AyLc+XeGaiE2htbdwbnYWMzVJID2JkZ3xutMqRVt7B48f66Ovz3vV19dXusaP4qFuTHshfLwKfO2u8HFCa/2f1rvm+PHj+uOPP96qKn5h6XSa8+fPMzAwsDx7lT+DVSqVYjGbXf9ipVB+EHHNIK5prxlSJLTsLBI+dqDy+I28HyLyK45zKKfgh4hcOUTg5NBOcd23NEyTaDRKLBYnkYhTnUgQi8VobGzk6aefpre3d7uv6r4hN5ntfg9/GJlMhkuXLjEwMMDFgQE+/+xzHP9nrwIhiuEa3Eid/5S8DjcUl5ZwIbYLrVH5JGZ2FiNzBzM7SyB3x1tHCm8s2KHDh3n82DH6+vo4evQosdh9FtTdOR7qHr4Xul3dAg6uOG4BxitUlw0RjUZ54YUXeOGFF9b8uuM4pNNpkslkeSXxVCpFMpksly3tz/vhJZmcILf4gNASCKHNIK4ZwDUCaCMAZgBt2mhz9b537O8by/vSl3lzuJFanLuOxSbRGlwHVSp664GUCijX23rH9+6rUhHlFjHcIkapAE7+vlPtGoZBVTRGLBajurqWhL9gYTweJxaLlffvPo5EIrIC+y5RVVXFiRMnOHHiBOBNnX7t2rXy68rVqwwNXabkeP/zlWnhhGspRWpxI/XeNlwt4wCF2GxuCSM3j5G9g5mZxVycxVqcK9/jDdOkra2Nw4ee59ChQ3R3d3Po0CFCoVCFK145e+GudB7oUUp1AGPAt4DvVLZKm8uyLKqrq6mu/mIzqCyFlvuFlWQySSaTIZ3JkElnyGQWyKazLC5m1121fBXDRFk2mDauEaBkWF5oMdYILav27RVhJ+B9oMofWWXSTP8QtLsiEBRhRWhYDhJ3BQfX+5qpnfI12il4AeQBrECAcDhCJBIhWlVFNFpLJBK5JzCsFSaqqqokRIhVbNvmyJEjHDlypFzmOA7Dw8Ncu3aNwcFBrl69yrVrg+SmLnsnKAWhOMVgAjdcgxuu9rahhDwIEuKLcl2M/ALG4jzG4hzG4jyB3DzkFsqfCXYwSFdXF4cP9dPT00NPTw/t7e3Ytl3hym8vuz58aK0dpdTrwNt4U+3+tdb6UoWrtS09amgBb82RXC5HNpslk8ms2q5VtnKbTnthJpv1QoxTXL9bSZlS5RCjzQAlZfkhJegFFctGm0FvsTgz6B/7LyvoDeCULgo7S3msQ94LDk7eDw7L+zj+cSmP4ToYrh80nEJ5EN+DBEMhIpEIkaoqotEY0aoqqqqqiEQiD7UNh8NEIhH5sBGbzrIsuru76e7uLpe5rsv4+DiDg4MMDQ0xPDzM4PUbTIx/iuv6s5wpBeEExWD1ciAJ1+AG4yCLVYq9TruofArTDxjG4hxWbgG1OF+eKVApxb7m/XQdOUpHRwft7e10d3fT0tKy3bu2bgu7fszHo9gN/YV3skKh8FCh5e5tKp0mlUqTTqfJZtLLH7TrUJYXTlzTpmQEVoQVfxzMyq1prwoy0pXhEWi93E2plPfGNJTyy4FhKUysCBDmUncmJ48uOfd9+4BtE62KEo1FicVixKLRVcHgYcJDOByWD47KkjEfm6RQKDA6Osrw8DBDQ0MMDQ1x/cYQk7cnllutleGHkjhuqBo3lPDCSSghM26J3cd1MHJ+S4a/tfJJVG4B3FL5tMamJro6O8sho6Ojg9bWVoLBYAUrv23JmA+xM9m2jW3bj9QCs0RrzeLiIul0mlQqteZ25X4ymSKZSpFK3SGbypDP5+7/DQzTH7hvUzJWtKoEwrh2BB2oQtsR3EDEm5lmN3Zx0BpKBYxCFlXMeIvdFbOoQmb1bEvuijEO93nYoZQiHKkiGq0iVhsnHqsnFov5A6hjq/aj0Wj5tXQsLQ1CrM+2bbq6uujq6lpVnsvluHnzJsPDwwwPDzMyMsLwyMjqlhJAhWIU7ThuOOEFEz+U3HfNGSG2AyePuTjvjcvwg0Ygn0TnkuVTlFI07Wumo/cwbW1ttLW1lUPGNpwVcMeT8CF2JaVU+Wn3oyzIUywW7wkoK0PL0vFS2UIySTKZYn5ujEIhf2997DBuIELJCuPaVWg/lCwFFdeObK8ZxbSLKi56gaKQQfmhYilcWM4iqpBZszUiGotTXZ0gHqshHo/fExLWChGxWIxIJIIhXT6E2FKhUIhDhw5x6NChVeXFYpGxsTFGRka4efOmF0qGR7g5ep1CfvkepwIhnGCCUiixHExCCXQwKl1bxdbR2vuMys1jLC5g5OYxcwtY+SS6sDyZTiBg03KwhY7247S1tdHa2kpbWxsHDhyQlowtJOFDiDUEAgFqamqoqan5QtdprUmn08zMzDAzM8P09PSq7eTUFNPTt0lNL9xzrTIttF2FY4XLLSd6Sz68tTe4upDFdLKYxax3s76rlcI0TWpq62hsbaCpsZH6+vryq6GhgYaGBurq6qQFQohdIBAI0N7eTnt7+6py13WZnp5mZGSkHEyG/f3kzNXyecqwcENxnOCKUBJO+IPd5U8P8YhcByOXXNVdysovYOSSq8b1VUVjtLe10db2ZLklo7W1laamJulauw3IHUCIDaSUKncR6ujoWPe8QqHA7OxsOZSsDCtT09NMTU0zN3sTt1Ra9z02UjAUor6+nqbGtnKYuHubSCSkZUKIPc4wDJqammhqaipPA7xkYWGB0dHRVa0lQ8MjTE0Mr5oNUYVi/gxc1ZSWunCFq73WXyEASkU/YHiDvs3cPFZ+AXKp8oMxpRT1DY10dHetasVobW39Ut22xeaT8CFEBdi2TXNzM83NzZWuihBCbIhEIkEikeDo0aOryvP5PGNjY+VA4oWSYUZHr6ya3VCFohTtRDmMLAUTLOkOs2uViqsChrE4TyC/gM6lyqeYpklLy0E6Op4sh4vW1lYOHjy4p9fK2MkkfAghhBBi0wSDQTo7O+ns7FxVXiqVmJiY8MeTeAPebwwNc/PmNYqTy4twqmBVea2SUrgGVxZQ3HnKC/HNYS7ewcjOESgsoHPp8ilWIMDBgwfp7DhRHvDd1tbG/v37sSz5We8m8tMUQgghxJbznmi30NLSwrPPPlsud12XycnJciDxQskQw8MrQkl5rZIa3EiNv6J7rTe74HaZuGMv0hpVzHqrfS/OYWTveAvxrVgjwwoEaGtto7PzMdrb22lra6O9vZ3m5mYZj7FHSPgQQgghxLZhGEa5W2p/f3+5fKml5Pr169y4cYPr169zbXCQybGh8jnKsnHKLST1lKrqcMM1MvPWZtCuNy4jM4OZmcVcnMPKzaGLy1PV1zc0cujxXjo7O+nq6qKzs5MDBw5IS8YeJz99IYQQQmx7K1tKnn/++XJ5Npv1Fk30Q8ng4CCD12+Qm7oMeDNvlSK1OEthpKrem3VLAsnD0xqVT2JmZrxXdgYre6c8w1QwGKKzs4Pu7uPloNHR0UEsFqtwxcV2JOFDCCGEEDtWJBLhyJEjHDlypFzmui7j4+NcuXKFy5cvc/nyZa5evUZ+6jMAlBnAidRRitRRqmqgFNuHtmUxuSWquIiZmsTITGMtBQ3HW98lELDp7unmsV97jsOHD3P48GFaWlqky5R4aBI+hBBCCLGrGIZRbiV56aWXAK/b1ujoKJcvX+bKlSt8fvky1wevUpy85F0UTlCoaqQU2+eHkeieGT+i8mnM9CRm6jZ2Zgqyc4DX2tTZ1cWv9f4Gvb29HD7srQAu3abElyG/PUIIIYTY9UzTLC+cePLkSQAcx2FwcJCBgQEuXrzIJxcHyMxcA7xZtgpVTZRi+3DizehQopLV31Aqn8JKTmCmbhPITHrrZwDhcIS+vl/n2LFjHDt2jJ6eHlk4Vmw4CR9CCCGE2JMsy6K3t5fe3l6++c1v4rouw8PDDAwMMDAwwIVffsL8yA3v5HCCQrwFp/ogpWgTGDuom5F2MdNTmPOj2MlbKL9lIxZP8MSJJzh27Bh9fX10dnZK9ymx6SR8CCGEEELgdddaWpPkzJkzaK0ZGxvj/PnzfPhP/8SFCxcoTV5CWTaF2H4viCRa0IFwpat+LyePtXALa34UOzWGLuYxTJO+vj6efeY7nDhxgtbWVtQe6Vomtg8JH0IIIYQQa1BKlceOnD17lmw2y4ULF/joo4/4fx98yPzQ/wVlUEy04NR34yQOVrZFRLuYC2MEZq4RWBgFt0QsnuCZF79Kf38/x48fJxqNVq5+QgBKa13pOmw7x48f1x9//HGlqyGEEHvNhjyClXu42Apaa65evcq7777LW2+/zfzcHCoQIl/bSbG+BzdSt2V1MRbnsWauEbpzHV3IEo3F+Npv/RYvvvgivb290pVKbJWHuodLy4cQQgghxBeklCpPNfvaa69x/vx53nzzLT748APsyc9wY/vI7X+cUqx502bNMlKThCY+wVwYwzAMnnrqKb7+9a/T399PIBDYlO8pxJcl4UMIIYQQ4kuwLIv+/n76+/tZWFjgF7/4Bf/jjTe4c+Ut3FgTuebHKcX3b1gIMVOTBMd/iZkcJxZP8K3f/V1OnjxJXd3WtbYI8agkfAghhBBCbJBEIsGrr77KqVOnePPNN/nbv/07Zq++jbIjYGzAqupao/MZ4olqvvO973H69GnC4W044F2IdUj4EEIIIYTYYMFgkDNnzvCNb3yDt956i88//3zD3rurq4uXX36ZUCi0Ye8pxFaR8CGEEEIIsUls2+bUqVOcOnWq0lURYlvYgPY/IYQQQgghhHgwCR9CCCGEEEKILSHhQwghhBBCCLElJHwIIYQQQgghtoSEDyGEEEIIIcSWkPAhhBBCCCGE2BIVCx9KqVeVUpeUUq5S6vhdX/tjpdSgUuqKUuprK8pP+mWDSqk/WlHeoZT6/0qpa0qpnyilbL886B8P+l9v36p/nxBCCCGEEGK1SrZ8/Ar4d8D7KwuVUo8B3wKOACeBHyulTKWUCfw58HXgMeDb/rkA/wX4oda6B5gDfscv/x1gTmvdDfzQP08IIYQQQghRARULH1rrz7XWV9b40mngDa11Xms9BAwCJ/zXoNb6hta6ALwBnFZKKeBF4O/96/8GOLPivf7G3/974CX/fCGEEEIIIcQW244rnB8APlpxfMsvAxi9q/wpoA6Y11o7a5x/YOkarbWjlFrwz5+5+5sqpb4LfNc/TCul1gpGQmx39azx+y3EDvGW1vrko1wo93CxS8g9XOxkD3UP39TwoZT638C+Nb70J1rrf1zvsjXKNGu30uj7nH+/97q3UOu/BP5ynToJsSMopT7WWh9/8JlC7C5yDxe7gdzDxV6wqeFDa/2bj3DZLeDgiuMWYNzfX6t8BqhWSll+68fK85fe65ZSygISwJ1HqJMQQgghhBDiS9qOU+3+DPiWP1NVB9AD/DNwHujxZ7ay8Qal/0xrrYF3gVf8638b+McV7/Xb/v4rwDv++UIIIYQQQogtVsmpds8qpW4B/cDPlVJvA2itLwE/BT4D3gJ+T2td8ls1XgfeBj4HfuqfC/CHwB8opQbxxnT8lV/+V0CdX/4HQHl6XiF2Kel2IoQQO5fcw8Wup6QhQAghhBBCCLEVtmO3KyGEEEIIIcQuJOFDCCGEEEIIsSUkfAghhBBCCCG2hIQPIYQQQgghxJaQ8CGEEEIIIYTYEhI+hBBCCCGEEFtCwocQQgghhBBiS/wrWDToS1TTdeUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "columns = [ ('transaction_count', 'box'), ('total_amount_usd', 'violin'),\n",
    "           ('median_amount_usd', 'violin'), ('has_email', 'count'), ('birth_year', 'violin'), ('kyc', 'count'), \n",
    "           ('country_name', 'count')]\n",
    "## Remove above 99th quantile of USD_TOTAL_AMOUNT\n",
    "quantile_99th = df_users[\"total_amount_usd\"].quantile(0.99)\n",
    "\n",
    "df_users_wo_outliers = df_users[df_users['total_amount_usd'] < quantile_99th]\n",
    "\n",
    "for column, plot_kind in columns:\n",
    "    users_by_column = df_users_wo_outliers[['id', column]].groupby([column]).count()\\\n",
    "                            .unstack()['id'].replace(np.nan, 0).sort_values(ascending=False)\n",
    "    top_ten_cols = None\n",
    "    if plot_kind == 'count':\n",
    "        df_fraudster_col = df_users_wo_outliers[df_users_wo_outliers['is_fraudster']==True][column].value_counts().sort_values(ascending=False)\n",
    "        df_normal_col = df_users_wo_outliers[df_users_wo_outliers['is_fraudster']==False][column].value_counts().sort_values(ascending=False)\n",
    "        top_ten_cols = df_fraudster_col.index[:10].union(df_normal_col.index[:10])\n",
    "    \n",
    "    \n",
    "    g = sns.catplot(y=column,\n",
    "             col=\"is_fraudster\", \n",
    "             data=df_users_wo_outliers, kind=plot_kind,\n",
    "             dodge=True,\n",
    "             order=top_ten_cols,\n",
    "             sharex=False if plot_kind=='count' else True,\n",
    "             orient='v',\n",
    "             height=4, aspect=1.4);\n",
    "    \n",
    "    #g.set_xticklabels(rotation=40)\n",
    "\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 806.4x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "columns = ['state', 'merchant_category', 'merchant_country_name', 'currency', 'entry_method', 'type', 'source']\n",
    "\n",
    "for column in columns:\n",
    "    users_by_column = df_transactions[['id', column]].groupby([column]).count()\\\n",
    "                            .unstack()['id'].replace(np.nan, 0).sort_values(ascending=False)\n",
    "    top_ten_cols = None\n",
    "    df_fraudster_col = df_transactions[df_transactions['is_fraudster']==True][column].value_counts().sort_values(ascending=False)\n",
    "    df_normal_col = df_transactions[df_transactions['is_fraudster']==False][column].value_counts().sort_values(ascending=False)\n",
    "    top_ten_cols = df_fraudster_col.index[:10].union(df_normal_col.index[:10], sort=False)\n",
    "   \n",
    "    g = sns.catplot(y=column,\n",
    "             col=\"is_fraudster\", \n",
    "             data=df_transactions, kind='count',\n",
    "             dodge=True,\n",
    "             order=top_ten_cols,\n",
    "             sharex=False,\n",
    "             sharey=False,\n",
    "             orient='v',\n",
    "             height=4, aspect=1.4);\n",
    "    plt.show()\n",
    "\n",
    "if False:\n",
    "    for column in columns:\n",
    "        transactions_by_column = df_transactions[['id', column]].groupby([column]).count()\\\n",
    "                                .unstack()['id'].replace(np.nan, 0).sort_values(ascending=False)\n",
    "\n",
    "        fraudster_by_column = df_transactions[['id', 'is_fraudster', column]].groupby([column, 'is_fraudster']).count()\\\n",
    "                                .unstack()[\"id\"].replace(np.nan, 0)\n",
    "        fig, axes = plt.subplots(1, 2)\n",
    "\n",
    "        fraudster_by_column_pos = fraudster_by_column.loc[:, False].sort_values(ascending=False)\n",
    "        fraudster_by_column_pos = merge_small_columns(fraudster_by_column_pos, 0.01)\n",
    "\n",
    "        fraudster_by_column_pos.plot.pie(ax=axes[0], figsize=(20, 20), autopct='%1.2f%%', pctdistance=0.8, labeldistance=1.1)\n",
    "\n",
    "        fraudster_by_column_neg = fraudster_by_column.loc[:, True].sort_values(ascending=False)\n",
    "        fraudster_by_column_neg = merge_small_columns(fraudster_by_column_neg, 0.01)\n",
    "\n",
    "        fraudster_by_column_neg.plot.pie(ax=axes[1], figsize=(20, 20), autopct='%1.2f%%', pctdistance=0.8, labeldistance=1.2)\n",
    "\n",
    "        [t.set_yticklabels(t.get_yticklabels(), rotation=90)  for t in axes]\n",
    "\n",
    "        axes[0].set_title('Normal user')\n",
    "        axes[1].set_title('Fraudster')\n",
    "        axes[0].set_ylabel(None)\n",
    "        axes[1].set_ylabel(None)\n",
    "        fig.suptitle('Chart for ' + column)\n",
    "        fig.subplots_adjust(top=1.4)\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preliminary observations\n",
    "\n",
    "There are very few fraudsters (238), which makes fraud detection task difficult.\n",
    "\n",
    "Fraudsters tend to heavily use GBP (91% of all transactions), while normal users tend to use both GBP (47%) and EUR (40%) in their transactions. Also they are mainly from UK, whereas normal users are more diverse in terms of the country where they are from.\n",
    "\n",
    "\n",
    "Fraudsters also tend to do bigger amount of transactions than normal users.\n",
    "\n",
    "Source seems interesting as well. MINOS is around 1.24% of all transactions in normal users. It is over 20% for fraudsters.  \n",
    "\n",
    "However these findings may not be conclusive and representative as we have very few positive,namely Fraudster, samples (298)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature engineering\n",
    "\n",
    "As we observed that currency and merchant_country seems to be considerably different in fraudsters, it may be wise to create features that describes the distribution of a user's currrency in their transactions. Likewise, we can also include the merchant country as one-hot encoding vector.\n",
    "\n",
    "Overall, the following features are derived:\n",
    "\n",
    "\n",
    "~~I first retrieved the currency rates from 2 websites, that gives me rates from all the currencies to USD in order to come up with a same unit for each transaction. Then I used USD currency as a reference for each transaction's amount. (AMOUNT_USD was not present in the transactions table despite it is described in the README.md)~~ (I reported this to HR guys, they uploaded new transaction table with amount_usd. However, my methods fetching the currency rates from the APIs are still present for imputting missing amount_usd values.)\n",
    "\n",
    "**MEDIAN_AMOUNT_USD**: Median amount among a user's transactions \n",
    "\n",
    "**TOTAL_AMOUNT_USD**: Total amount of a user's transactions\n",
    "\n",
    "**TRANSACTION_COUNT**: Number of a user's transactions\n",
    "\n",
    "**MEAN_TRANSACTION_PERIOD**: This is average time difference between two transaction of a user. \n",
    "\n",
    "**CURRENCY[]**: A vector that sums up to 1. Each dimension represents the probability that a user is made transaction with a given currency.\n",
    "\n",
    "**MERCHANT_COUNTRY[]**: A vector that sums up to 1. Each dimension represents the probability that a user is made transaction to a given merchant country.\n",
    "\n",
    "**SOURCE[]**: A vector that sums up to 1. Each dimension represents the probability that a user is made transaction from a given source.\n",
    "\n",
    "**STATE[] (from transactions table)**: A vector that sums up to 1. Each dimension represents the probability that a user's transaction state is the corresponding state. \n",
    "\n",
    "I did not include **STATE** variable from df_users as it is 100% correlated with IS_FRAUDSTER. This led me to believe that they are indeed created from each other. If I include it to my model, I get 100% accuracy in the test set. So, I believe that you would remove it but you just forgot. In addition, my model is intended to take an action and change the state variable. Therefore, it wouldn't make sense to use my future output as my input. \n",
    "\n",
    "\n",
    "For one-hot encoding vectors, I filtered dimensions by choosing only the features that consists more than 0.1% of transactions. As an example, Grenada will not be included to MERCHANT_COUNTRY vector as it is only present in 33 transactions ( <0.0001%), as these features have very low variance and will probably ineffective in prediction. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create and load features into database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "!python3 -W ignore ../code/features.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Retrieving features from csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_dataset = pd.read_csv('../data/processed_dataset.csv').set_index('user_id')\n",
    "processed_labels = pd.read_csv('../data/processed_labels.csv', header=None).set_index(0).loc[:, 1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Retrieve features from database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "processed_dataset, processed_labels = get_features(engine)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model training and evaluation\n",
    "\n",
    "I run XGBClassifier for this challenge. The reason is that it handles missing features, feature selection and regularization for myself. Additionally, it usually performs well in structured data. \n",
    "\n",
    "In order to evaluate the model, I initially split the dataset (stratified) into train (75%) and test sets (25%) . Afterwards, I did hyperparameter tuning with stratified kfold cross validation where `n_splits=3` \n",
    "\n",
    "After hyperparameter tuning with `Randomized Grid Search`, I used the hold out test set to evaluate the performance of the trained optimal model. \n",
    "\n",
    "As classes are unbalanced in the dataset due to the nature of the problem, I used F1 score and Matthews correlation coefficient that takes both false negatives and false positives into account. Therefore, metrics are appropriate for this domain, where the classes are highly unbalanced.\n",
    "\n",
    "I repeated the aforementioned procedure **3 times** and evaluated the means and standard deviations of the scores.\n",
    "\n",
    "\n",
    "Matthews correlation coefficient turned out to be around 0.76±0.03 which is actually much better than a poor model :). \n",
    "\n",
    "### Discussion\n",
    "\n",
    "We could make the model more advanced by using stacking. Sometimes, it improves the performance. However, this is the final week at my university, so I don't have time to do this. ( I can do more analysis after next week if you'd like me to do)\n",
    "\n",
    "One **assumption** I made is that the data does not depend on time (stationarity). Unfortunately, I didn't get any evidence that proves or rejects this assumption during my analysis."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Obtain 3 XGB Models with hyperparameter tuning using different train test splits\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 80 candidates, totalling 240 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=10)]: Using backend LokyBackend with 10 concurrent workers.\n",
      "[Parallel(n_jobs=10)]: Done  21 tasks      | elapsed:    5.2s\n",
      "[Parallel(n_jobs=10)]: Done 142 tasks      | elapsed:   20.6s\n",
      "[Parallel(n_jobs=10)]: Done 240 out of 240 | elapsed:   31.6s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 80 candidates, totalling 240 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=10)]: Using backend LokyBackend with 10 concurrent workers.\n",
      "[Parallel(n_jobs=10)]: Done  21 tasks      | elapsed:    4.2s\n",
      "[Parallel(n_jobs=10)]: Done 142 tasks      | elapsed:   19.4s\n",
      "[Parallel(n_jobs=10)]: Done 240 out of 240 | elapsed:   30.2s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 80 candidates, totalling 240 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=10)]: Using backend LokyBackend with 10 concurrent workers.\n",
      "[Parallel(n_jobs=10)]: Done  21 tasks      | elapsed:    4.2s\n",
      "[Parallel(n_jobs=10)]: Done 142 tasks      | elapsed:   19.5s\n",
      "[Parallel(n_jobs=10)]: Done 240 out of 240 | elapsed:   30.6s finished\n"
     ]
    }
   ],
   "source": [
    "from train import train_xgb_model\n",
    "from sklearn.model_selection import train_test_split\n",
    "random_states = [42, 52, 62]\n",
    "scores = []\n",
    "\n",
    "samples_with_labels = ~processed_labels.isna()\n",
    "processed_dataset = processed_dataset[samples_with_labels]\n",
    "processed_labels = processed_labels[samples_with_labels]\n",
    "\n",
    "for random_state in random_states:\n",
    "    \n",
    "    train_set, test_set, train_labels, test_labels = train_test_split(processed_dataset, processed_labels, test_size=0.25,\n",
    "                                                                 stratify=processed_labels, random_state=random_state)\n",
    "    clf = train_xgb_model(train_set, train_labels)\n",
    "    scores.append(get_metric_results(clf, test_set, test_labels))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluate the performance using relevant metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n"
     ]
    },
    {
     "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>metric</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>avg_precision</td>\n",
       "      <td>0.84±0.02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>f1_score</td>\n",
       "      <td>0.77±0.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>matthews</td>\n",
       "      <td>0.76±0.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>mse</td>\n",
       "      <td>0.05±0.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>roc_auc</td>\n",
       "      <td>0.98±0.01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          metric      score\n",
       "0  avg_precision  0.84±0.02\n",
       "1       f1_score  0.77±0.03\n",
       "2       matthews  0.76±0.03\n",
       "3            mse  0.05±0.01\n",
       "4        roc_auc  0.98±0.01"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import * \n",
    "from sklearn.metrics import confusion_matrix\n",
    "fig = plt.figure()\n",
    "plot_confusion_matrix(fig, confusion_matrix(test_labels, clf.predict(test_set)), classes=['Normal', 'Fraudster'])\n",
    "df_scores = pd.DataFrame(scores)\n",
    "df_scores = pd.concat([df_scores.agg(lambda x: str(round(x.mean(), 2)) + '±' + str(round(x.std(), 2)))], axis=0)\n",
    "df_scores = pd.DataFrame({'metric': df_scores.index, 'score': df_scores.values})\n",
    "\n",
    "df_scores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train a model and save it to artifacts/model.pkl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 80 candidates, totalling 240 fits\n",
      "[Parallel(n_jobs=10)]: Using backend LokyBackend with 10 concurrent workers.\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.01, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.01, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=100, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.01, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=100, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=100, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=6, learning_rate=0.01, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.6 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=6, learning_rate=0.01, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.6, total=   0.8s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=6, learning_rate=0.01, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=6, learning_rate=0.01, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.6, total=   0.6s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=6, learning_rate=0.01, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=6, learning_rate=0.01, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.6, total=   0.6s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=100, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=0.7, total=   2.5s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=100, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=0.7, total=   2.6s\n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=100, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=0.7, total=   2.6s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.5 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.001, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.9, total=   2.7s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.001, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.9, total=   2.8s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.001, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.5, total=   0.5s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.5, total=   0.4s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.5, total=   0.5s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.01, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.7, total=   3.2s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.9, total=   3.1s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.01, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.7, total=   3.3s\n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.01, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.7, total=   3.4s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.8 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.9, total=   1.1s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.8, total=   1.7s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.8, total=   1.7s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.8, total=   1.7s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.6 \n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.9, total=   1.0s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.9, total=   1.1s\n",
      "[Parallel(n_jobs=10)]: Done  21 tasks      | elapsed:    6.9s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.9, total=   3.7s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.9, total=   3.9s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.001, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.9, total=   4.2s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=100, min_child_weight=1.0, max_depth=10, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.7, colsample_bylevel=0.9, total=   3.9s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.6, total=   2.0s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=100, min_child_weight=1.0, max_depth=10, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=100, min_child_weight=1.0, max_depth=10, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.001, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.9, total=   4.6s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.001, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.9, total=   4.5s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.8 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.6, total=   2.3s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.8 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.6, total=   2.1s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=100, min_child_weight=1.0, max_depth=10, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   1.6s\n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=100, min_child_weight=1.0, max_depth=10, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   1.8s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6, total=   1.1s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=100, min_child_weight=1.0, max_depth=10, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   1.7s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.8, total=   1.9s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.8, total=   1.8s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.8, total=   1.9s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=1.0, total=   3.2s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.4, colsample_bylevel=0.5, total=   0.8s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0, colsample_bytree=0.7, colsample_bylevel=0.7 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.4, colsample_bylevel=0.5, total=   0.8s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0, colsample_bytree=0.7, colsample_bylevel=0.7 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6, total=   1.1s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=1.0, total=   3.1s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0, colsample_bytree=0.7, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=100, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0.25, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6, total=   1.1s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.4, colsample_bylevel=0.5, total=   0.8s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=100, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0.25, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=100, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0.25, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=1.0, total=   3.2s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.9 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0, colsample_bytree=0.7, colsample_bylevel=0.7, total=   1.4s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0, colsample_bytree=0.7, colsample_bylevel=0.7, total=   1.4s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   0.4s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0, colsample_bytree=0.7, colsample_bylevel=0.7, total=   1.5s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=100, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0.25, colsample_bytree=1.0, colsample_bylevel=0.5, total=   1.5s\n",
      "/home/ibrahim/miniconda3/envs/dta_pred/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=100, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0.25, colsample_bytree=1.0, colsample_bylevel=0.5, total=   1.5s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   0.4s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=100, min_child_weight=10.0, max_depth=15, learning_rate=0.01, gamma=0.25, colsample_bytree=0.8, colsample_bylevel=0.9 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=100, min_child_weight=10.0, max_depth=15, learning_rate=0.01, gamma=0.25, colsample_bytree=0.8, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=100, min_child_weight=3.0, max_depth=15, learning_rate=0, gamma=0.25, colsample_bytree=1.0, colsample_bylevel=0.5, total=   1.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=100, min_child_weight=10.0, max_depth=15, learning_rate=0.01, gamma=0.25, colsample_bytree=0.8, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   0.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.9, total=   2.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=100, min_child_weight=10.0, max_depth=15, learning_rate=0.01, gamma=0.25, colsample_bytree=0.8, colsample_bylevel=0.9, total=   1.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=100, min_child_weight=10.0, max_depth=15, learning_rate=0.01, gamma=0.25, colsample_bytree=0.8, colsample_bylevel=0.9, total=   1.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=7.0, max_depth=6, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=100, min_child_weight=10.0, max_depth=15, learning_rate=0.01, gamma=0.25, colsample_bytree=0.8, colsample_bylevel=0.9, total=   1.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=7.0, max_depth=6, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   1.7s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=7.0, max_depth=6, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.9, total=   2.3s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.9, total=   2.4s\n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   1.7s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.9 \n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   1.8s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=70, min_child_weight=7.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=7.0, max_depth=6, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.9, total=   0.5s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=70, min_child_weight=7.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=7.0, max_depth=6, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.9, total=   0.5s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=70, min_child_weight=7.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=7.0, max_depth=6, learning_rate=3, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=0.9, total=   0.5s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=70, min_child_weight=7.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.7, total=   0.9s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=1.0 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.7, total=   2.7s\n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=70, min_child_weight=7.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.7, total=   0.9s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0.01, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=70, min_child_weight=7.0, max_depth=20, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.7, total=   1.0s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0.01, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=1.0, total=   1.1s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0.01, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0.01, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.7, total=   0.8s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0.01, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.7, total=   0.8s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=100.0, n_estimators=70, min_child_weight=3.0, max_depth=15, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=100.0, n_estimators=70, min_child_weight=3.0, max_depth=15, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.7, total=   2.6s\n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=1.0, total=   1.1s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=100.0, n_estimators=70, min_child_weight=3.0, max_depth=15, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=5.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=7.0, max_depth=15, learning_rate=0.01, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.7, total=   0.8s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=5.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=1, reg_lambda=1.0, n_estimators=100, min_child_weight=10.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=1.0, total=   1.2s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=5.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=0.7, total=   2.8s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=5.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   0.4s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=5.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   0.3s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.9 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=5.0, max_depth=10, learning_rate=3, gamma=1.0, colsample_bytree=0.8, colsample_bylevel=0.6, total=   0.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=100.0, n_estimators=100, min_child_weight=7.0, max_depth=20, learning_rate=0.2, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.9, total=   0.7s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=100.0, n_estimators=100, min_child_weight=7.0, max_depth=20, learning_rate=0.2, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.9, total=   1.0s\n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.9, total=   1.2s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=100.0, n_estimators=100, min_child_weight=7.0, max_depth=20, learning_rate=0.2, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.9, total=   4.0s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.9, total=   4.1s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.9, total=   4.2s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.6 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   0.9s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=100.0, n_estimators=70, min_child_weight=3.0, max_depth=15, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=1.0, total=   2.8s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=100.0, n_estimators=70, min_child_weight=3.0, max_depth=15, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=1.0, total=   2.8s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   0.9s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=100.0, n_estimators=70, min_child_weight=3.0, max_depth=15, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=1.0, total=   2.7s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   0.9s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=100.0, n_estimators=100, min_child_weight=7.0, max_depth=20, learning_rate=0.2, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.5, total=   2.2s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=100.0, n_estimators=100, min_child_weight=7.0, max_depth=20, learning_rate=0.2, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.5, total=   2.2s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=100.0, n_estimators=100, min_child_weight=7.0, max_depth=20, learning_rate=0.2, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.5, total=   2.3s\n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=1.0, total=   1.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=1.0, total=   1.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=1.0, total=   1.4s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.8 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=1.0, total=   1.2s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.8 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=1.0, total=   1.2s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.8 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=100.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=0.5 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=0.5, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.6, total=   3.8s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=0.5, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=0.5, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.6, total=   3.6s\n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.001, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=0.1, n_estimators=150, min_child_weight=10.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.6, total=   3.9s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.8, total=   1.8s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.8, total=   1.9s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.8, total=   1.8s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   1.2s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   1.2s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=100, min_child_weight=3.0, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   1.2s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=0.5, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.9, total=   2.4s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=0.5, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.9, total=   2.3s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.7, total=   1.4s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=150, min_child_weight=0.5, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.9, total=   2.3s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.7, total=   1.4s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.7, total=   1.4s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   1.7s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   1.6s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=6, learning_rate=0.01, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   1.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   4.0s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   3.8s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=1, reg_lambda=0.1, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   4.0s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   1.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.7, total=   3.0s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.7, total=   3.4s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=10.0, max_depth=15, learning_rate=0, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0, gamma=0.5, colsample_bytree=0.9, colsample_bylevel=0.7, total=   3.1s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=10.0, max_depth=15, learning_rate=0, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   1.5s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=10.0, max_depth=15, learning_rate=0, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   1.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   3.6s\n",
      "[Parallel(n_jobs=10)]: Done 142 tasks      | elapsed:   28.9s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=10.0, max_depth=15, learning_rate=0, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   0.9s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   3.3s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=10.0, max_depth=15, learning_rate=0, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   0.8s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=10.0, max_depth=15, learning_rate=0, gamma=1.0, colsample_bytree=0.6, colsample_bylevel=1.0, total=   0.8s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   3.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   2.9s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.5, total=   0.7s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   2.6s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.5 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=0.1, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.5, total=   2.7s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   1.6s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   1.7s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.5, total=   0.7s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.5, total=   0.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   2.3s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   2.2s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=15, learning_rate=0.001, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.4, total=   1.6s\n",
      "[CV] subsample=0.9, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=5.0, n_estimators=150, min_child_weight=3.0, max_depth=10, learning_rate=0.2, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   2.2s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.6, total=   1.5s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=1.0, total=   2.6s\n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=1.0, total=   2.7s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.6 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.5, total=   2.9s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.5, total=   3.1s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0.25, colsample_bytree=0.9, colsample_bylevel=1.0, total=   2.7s\n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=1.0, total=   2.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=1.0, total=   2.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.9, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=1.0, colsample_bytree=0.7, colsample_bylevel=1.0, total=   2.7s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=1.0, max_depth=20, learning_rate=0.001, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.5, total=   3.0s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.6, total=   1.4s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.7, total=   0.4s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.7, total=   0.5s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.4, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.6, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=10.0, n_estimators=100, min_child_weight=0.5, max_depth=15, learning_rate=3, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.6, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=5.0, n_estimators=70, min_child_weight=3.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.8, colsample_bylevel=0.7, total=   0.4s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.6, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=1.0, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.01, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.6, total=   1.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=70, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6, total=   1.5s\n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6, total=   1.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=70, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=1.0, colsample_bylevel=0.6, total=   1.6s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=70, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.4, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.4, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.9 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=7.0, max_depth=10, learning_rate=0.001, gamma=0.5, colsample_bytree=0.4, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=70, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=70, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=70, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=1.0, colsample_bytree=0.5, colsample_bylevel=1.0, total=   1.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.9, total=   1.8s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.9, total=   1.9s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=4, reg_lambda=10.0, n_estimators=70, min_child_weight=5.0, max_depth=6, learning_rate=0.1, gamma=1.0, colsample_bytree=0.9, colsample_bylevel=0.9, total=   1.9s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=100, min_child_weight=5.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.5, colsample_bylevel=0.6 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.6, colsample_bylevel=0.6, total=   3.0s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=100, min_child_weight=5.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.5, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=100, min_child_weight=5.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.5, colsample_bylevel=0.6, total=   0.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=100, min_child_weight=5.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.5, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.7, total=   2.2s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.6, colsample_bylevel=0.6, total=   3.2s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=2, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=0, gamma=0.5, colsample_bytree=0.6, colsample_bylevel=0.6, total=   3.3s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=100, min_child_weight=5.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.5, colsample_bylevel=0.6, total=   0.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=100, min_child_weight=5.0, max_depth=6, learning_rate=3, gamma=0.5, colsample_bytree=0.5, colsample_bylevel=0.6, total=   0.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.5, total=   1.5s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.7 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.7, total=   2.1s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.5, total=   1.5s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=20, learning_rate=0.1, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.7, total=   2.2s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=100, min_child_weight=3.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.6, colsample_bylevel=0.5, total=   1.6s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=150, min_child_weight=10.0, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.4, total=   0.5s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=150, min_child_weight=10.0, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.7, total=   1.3s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.4, total=   0.5s\n",
      "[CV] subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=150, min_child_weight=10.0, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.6 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=70, min_child_weight=10.0, max_depth=6, learning_rate=0.1, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.4, total=   0.5s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.7, total=   1.3s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.4, colsample_bylevel=1.0 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=0.1, n_estimators=70, min_child_weight=0.5, max_depth=10, learning_rate=0.001, gamma=1.0, colsample_bytree=0.4, colsample_bylevel=0.7, total=   1.4s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.7, total=   2.0s\n",
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.7, total=   2.1s\n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=1.0, max_depth=6, learning_rate=0.1, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.7, total=   2.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.8 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=150, min_child_weight=10.0, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=150, min_child_weight=10.0, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.5, total=   0.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=100, min_child_weight=1.0, max_depth=6, learning_rate=0.001, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.7, silent=False, scale_pos_weight=4, reg_lambda=50.0, n_estimators=150, min_child_weight=10.0, max_depth=15, learning_rate=0.2, gamma=0.25, colsample_bytree=0.4, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=100, min_child_weight=1.0, max_depth=6, learning_rate=0.001, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.5, total=   0.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=100, min_child_weight=1.0, max_depth=6, learning_rate=0.001, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.8 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.4, colsample_bylevel=1.0, total=   1.5s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=1.0, n_estimators=150, min_child_weight=0.5, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.6, colsample_bylevel=0.5, total=   0.6s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.4, colsample_bylevel=1.0, total=   1.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=5.0, n_estimators=150, min_child_weight=7.0, max_depth=6, learning_rate=0.2, gamma=0, colsample_bytree=0.4, colsample_bylevel=1.0, total=   1.5s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.8, total=   1.6s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.8, total=   1.6s\n",
      "[CV] subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.8, silent=False, scale_pos_weight=2, reg_lambda=0.1, n_estimators=70, min_child_weight=1.0, max_depth=10, learning_rate=0.001, gamma=0.25, colsample_bytree=0.5, colsample_bylevel=0.8, total=   1.6s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.2, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   0.9s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.2, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.6 \n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.2, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.6 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=100.0, n_estimators=150, min_child_weight=5.0, max_depth=10, learning_rate=0, gamma=0.5, colsample_bytree=0.7, colsample_bylevel=0.5, total=   1.0s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=100, min_child_weight=1.0, max_depth=6, learning_rate=0.001, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.8, total=   1.8s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=100, min_child_weight=1.0, max_depth=6, learning_rate=0.001, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.8, total=   1.9s\n",
      "[CV] subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.4 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.4, total=   0.8s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.5 \n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   1.8s\n",
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=1, reg_lambda=50.0, n_estimators=100, min_child_weight=1.0, max_depth=6, learning_rate=0.001, gamma=0.5, colsample_bytree=1.0, colsample_bylevel=0.8, total=   2.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.5 \n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.2, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   1.8s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.2, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=32.308724832214764, reg_lambda=50.0, n_estimators=70, min_child_weight=1.0, max_depth=15, learning_rate=0.2, gamma=0, colsample_bytree=0.5, colsample_bylevel=0.6, total=   1.3s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.4, total=   0.8s\n",
      "[CV]  subsample=1.0, silent=False, scale_pos_weight=1, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=15, learning_rate=0.1, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.4, total=   1.7s\n",
      "[CV]  subsample=0.6, silent=False, scale_pos_weight=4, reg_lambda=5.0, n_estimators=70, min_child_weight=10.0, max_depth=20, learning_rate=0, gamma=0, colsample_bytree=0.8, colsample_bylevel=0.4, total=   0.8s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.5, total=   0.7s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.5, total=   0.7s\n",
      "[CV]  subsample=0.5, silent=False, scale_pos_weight=2, reg_lambda=10.0, n_estimators=150, min_child_weight=5.0, max_depth=20, learning_rate=3, gamma=0, colsample_bytree=0.9, colsample_bylevel=0.5, total=   1.2s\n",
      "[Parallel(n_jobs=10)]: Done 240 out of 240 | elapsed:   44.5s finished\n"
     ]
    }
   ],
   "source": [
    "!python3 -W ignore ../code/train.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Patrol.py\n",
    "\n",
    "It may be wise to lock user when we are highly certain that the user is a fraudster. To do this, we may aim to\n",
    "maximize our utility. I defined a cost that denotes the loss of company when we didn't lock a fraudster but only\n",
    "alert the agent and another cost that denotes the loss of company when we locked a normal user despite that\n",
    "they are innocent. Hereby, the algorithm determines the action based on the losses given by the company.\n",
    "\n",
    "Costs can be updated from misc/patrol_config.yaml\n",
    "\n",
    "\n",
    "As a summary here is the utilities\n",
    "\n",
    "`action_costs = {\n",
    "    'LOCK_USER': prob_being_normal * costs['cost_of_false_negative'],\n",
    "    'ALERT_AGENT': prob_being_fraudster * costs['cost_of_false_negative'],\n",
    "    'NONE': prob_being_fraudster * costs['cost_of_false_positive']\n",
    "}`\n",
    "\n",
    "The action with the minimum cost is gonna be returned from the patrol method.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'LOCK_USER'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from patrol import patrol\n",
    "\n",
    "patrol('2707dd70-86d3-4823-ad0f-91a340bccb88')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}