experimento3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "4b9ee2a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "#importar librerias \n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "aa9b05f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv(\"C:/Users/jabel/OneDrive - usach.cl/Inv/UValpo/aprendizaje de maquina/trabajo_aprendizaje maquina/train.tsv\", delimiter = \"\\t\", quoting = 3)\n",
    "\n",
    "#dataset1 = pd.read_csv(\"C:/Users/jabel/OneDrive - usach.cl/Inv/UValpo/aprendizaje de maquina/trabajo_aprendizaje maquina/train.tsv\", delimiter = \"\\t\", quoting = 3)\n",
    "#dataset2  = pd.read_csv(\"C:/Users/jabel/OneDrive - usach.cl/Inv/UValpo/aprendizaje de maquina/trabajo_aprendizaje maquina/dev.tsv\", delimiter = \"\\t\", quoting = 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ec5a4055",
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "import nltk\n",
    "#nltk.download('stopwords')\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "corpus = []\n",
    "for i in range(0, 5886):\n",
    "    review = re.sub('[^a-zA-Z]', ' ', dataset['tweet'][i])\n",
    "    review = review.lower()\n",
    "    review = review.split()\n",
    "    ps = PorterStemmer()\n",
    "    review = [ps.stem(word) for word in review if not word in set(stopwords.words('spanish'))]\n",
    "    review = ' '.join(review)\n",
    "    corpus.append(review)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "44c18a29",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "cv = CountVectorizer(max_features = 1500)\n",
    "x = cv.fit_transform(corpus).toarray()\n",
    "y = dataset.iloc[:, 2].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8745b071",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.2, random_state = 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5324bcf",
   "metadata": {},
   "source": [
    "La matriz de confusión (confusion matrix ) presenta en una tabla una visión gráfica de los errores cometidos por el modelo de clasificación. Se trata de un modelo gráfico para visualizar el nivel de acierto de un modelo de predicción. Esta matriz indica el número de instancias correcta e incorrectamente clasificadas. Los parámetros que nos indica son: \n",
    "    \n",
    "        * Verdadero positivo (True Positive, TP): número de clasificaciones correctas en la clase positiva (P).\n",
    "        * Verdadero negativo (True Negative, TN): número de clasificaciones correctas en la clase negativa (N).\n",
    "        * Falso negativo (False Negative, FN): número de clasificaciones incorrectas de clase positiva clasificada como negativa.\n",
    "        * Falso positivo (False Positive, FP): número de clasificaciones incorrectas de clase negativa clasificada como positiva.\n",
    "\n",
    "\n",
    "* La matriz de confusión se puede extender a más de dos clases de forma natural. Esta representación nos permite identificar de forma rápida el número de instancias correctamente clasificadas, que se corresponde con la diagonal de la tabla. en el caso de problemas de clasificación que impliquen más de dos clases, se puede realizar la evaluación de dos formas distintas: \n",
    "  \n",
    "     * 1-vs-1: midiendo la capacidad de discriminar entre instancias de una clase, considerada la clase positiva, frente a las      instancias de otra clase, consideradas negativas.\n",
    "     * 1-vs-all: midiendo la capacidad de discriminar entre instancias de una clase, considerada la clase positiva, frente a        las instancias de las demás clases, consideradas negativas.\n",
    "  \n",
    " \n",
    "* El error de clasificación (misclassification error, ERR) es la suma de la predicciones incorrectas sobre el número total de predicciones. \n",
    "  $$ ERR = \\dfrac{FP+FN}{FP+FN+TP+TN}$$\n",
    "  \n",
    "* La exactitud (accuracy, ACC) es el número de predicciones correctas sobre el número total de predicciones.\n",
    "  $$ ACC= 1- ERR$$\n",
    "  \n",
    "* Tasa de verdaderos positivos (True Positive Rate, TPR)\n",
    "  $$ TPR = \\dfrac{TP}{FN+TP}$$\n",
    "  \n",
    "* Tasa de verdaderos negativos (False Positive Rate, FPR)\n",
    "  $$ FPR =  \\dfrac{FP}{FP+TN}$$\n",
    "      \n",
    "* La precisión (precision, PRE) mide el rendimiento relacionado con las tasas de verdaderos positivos y negativos.\n",
    "  $$ PRE = \\dfrac{TP}{TP+FP}$$\n",
    "  \n",
    "* El recall (recall, REC) y la sensibilidad (sensitivity, SEN) se corresponden con la tasa de verdaderos positivos (TPR), mientras que la especificidad (specificity, SPE) se define como la tasa de instancias correctamente clasificadas como negativas respecto a todas las instancias negativas.\n",
    "  $$ REC= SEN = TPR = \\dfrac{TP}{FN+TP}$$\n",
    "  $$ SPE = \\dfrac{TN}{TN+PP}= 1- FPR$$\n",
    "  \n",
    "* F1 (F1 score) es una combianción de la la precisión y el recall.\n",
    "  $$F1 = 2 \\dfrac{(PRE)(REC)}{(PRE)(REC)} $$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58a52588",
   "metadata": {},
   "source": [
    "### K- Nearest Neighbour (KNN). \n",
    "Es un algoritmo de aprendizaje supervisado de clasificación, de modo que a partir de un juego de datos de entrenamiento su objetivo será clasificar correctamente todas las instancias nuevas. El juego de datos típico de este tipo de algoritmos está formado por varios atributos descriptivos y un solo atributo objetivo, también llamado clase. En contraste con otros algoritmos de aprendizaje supervisado, el algoritmo kNN no genera un modelo fruto del aprendizaje con datos de entrenamiento, sino que el aprendizaje sucede en el mismo momento en el que se pide clasificar una nueva instancia. A este tipo de algoritmos se les llama métodos de aprendizaje perezoso o \\textit{lazy learning methods}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "790f67d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[  9   1   0  10 103   4   0]\n",
      " [  2   1   0   1  20   0   0]\n",
      " [  0   1   1   0  12   0   0]\n",
      " [  4   0   1  33 217   0   0]\n",
      " [ 18   0   5  18 528   1   0]\n",
      " [  6   0   1   4 116   9   0]\n",
      " [  2   0   0   5  45   0   0]]\n",
      "accuracy score =  0.4932088285229202\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.22      0.07      0.11       127\n",
      "    disgust        0.33      0.04      0.07        24\n",
      "       fear        0.12      0.07      0.09        14\n",
      "        joy        0.46      0.13      0.20       255\n",
      "     others        0.51      0.93      0.66       570\n",
      "    sadness        0.64      0.07      0.12       136\n",
      "   surprise        0.00      0.00      0.00        52\n",
      "\n",
      "    accuracy                           0.49      1178\n",
      "   macro avg       0.33      0.19      0.18      1178\n",
      "weighted avg       0.45      0.49      0.39      1178\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x20a6df1b430>"
      ]
     },
     "execution_count": 7,
     "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": [
    "#KNN\n",
    "#training model\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "knn = KNeighborsClassifier(n_neighbors = 5, metric = 'minkowski',p = 2)\n",
    "knn.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = knn.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "knna = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(knn,x_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7ffa103",
   "metadata": {},
   "source": [
    "## Maquina de Soporte Vectoriales.\n",
    "Las SVM son modelos de aprendizaje supervisado asociados a algoritmos de aprendizaje que inspeccionan datos utilizados para la clasificación y el análisis de regresión;  tanto lineales como no lineales. La hipótesis tradicional se basa en lo que se conoce como minimización del riesgo empírico (\\textit{empirical risk minimization}) mientras que el enfoque de las SVM se basa en la minimización del riesgo estructural (\\textit{structural risk minimization}), de modo que lo que se busca es construir modelos que estructuralmente tengan poco riesgo de cometer errores ante clasificaciones futuras \\cite{mt5}.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c0ee91a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\_base.py:255: ConvergenceWarning: Solver terminated early (max_iter=1000).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[ 59   7   0  12  33  11   5]\n",
      " [ 12   4   0   1   5   0   2]\n",
      " [  1   1   3   1   4   1   3]\n",
      " [ 12   5   1 121  91   8  17]\n",
      " [ 54  13   8 117 327  22  29]\n",
      " [  9   2   2  10  22  88   3]\n",
      " [  4   0   2  14  17   2  13]]\n",
      "accuracy score =  0.5220713073005093\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.39      0.46      0.42       127\n",
      "    disgust        0.12      0.17      0.14        24\n",
      "       fear        0.19      0.21      0.20        14\n",
      "        joy        0.44      0.47      0.46       255\n",
      "     others        0.66      0.57      0.61       570\n",
      "    sadness        0.67      0.65      0.66       136\n",
      "   surprise        0.18      0.25      0.21        52\n",
      "\n",
      "    accuracy                           0.52      1178\n",
      "   macro avg       0.38      0.40      0.39      1178\n",
      "weighted avg       0.54      0.52      0.53      1178\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x21c51ba1d00>"
      ]
     },
     "execution_count": 6,
     "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": [
    "#SVM\n",
    "#training model\n",
    "from sklearn.svm import SVC\n",
    "svc = SVC(kernel = 'linear',random_state = 0,  class_weight = 'balanced', decision_function_shape='ovr', max_iter = 1000)\n",
    "svc.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = svc.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "sva =accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(svc,x_test,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "afc234cd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[ 21   0   0   3  99   4   0]\n",
      " [  2   0   0   0  22   0   0]\n",
      " [  0   0   0   0  14   0   0]\n",
      " [  3   0   0  68 183   1   0]\n",
      " [  2   0   0  16 549   3   0]\n",
      " [  0   0   0   1  63  72   0]\n",
      " [  1   0   0   7  44   0   0]]\n",
      "accuracy score =  0.6027164685908319\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.72      0.17      0.27       127\n",
      "    disgust        0.00      0.00      0.00        24\n",
      "       fear        0.00      0.00      0.00        14\n",
      "        joy        0.72      0.27      0.39       255\n",
      "     others        0.56      0.96      0.71       570\n",
      "    sadness        0.90      0.53      0.67       136\n",
      "   surprise        0.00      0.00      0.00        52\n",
      "\n",
      "    accuracy                           0.60      1178\n",
      "   macro avg       0.41      0.27      0.29      1178\n",
      "weighted avg       0.61      0.60      0.53      1178\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x20a6d47d820>"
      ]
     },
     "execution_count": 9,
     "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": [
    "#Kernel SVM\n",
    "#training model\n",
    "from sklearn.svm import SVC\n",
    "svc = SVC(kernel = 'rbf', random_state = 0)\n",
    "svc.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = svc.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "sva2 = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(svc,x_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a5f465ac",
   "metadata": {},
   "source": [
    "## Naive Bayes.\n",
    "Se basa en el concepto de probabilidad condicional y busca maximizar la verosimilitud del modelo, es decir, otorgar mayor importancia a aquellos eventos que son realmente relevantes en el juego de datos.En general, los métodos estadísticos suelen estimar un conjunto de parámetros probabilísticos, que expresan la probabilidad condicionada de cada clase dadas las propiedades de un ejemplo (descrito en forma de atributos).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "8dfec5d0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[ 38   8   3  18  34  19   7]\n",
      " [  9   2   0   3   4   4   2]\n",
      " [  4   1   3   1   2   1   2]\n",
      " [ 25   7   3  93  49  43  35]\n",
      " [103  23  26 136 120 107  55]\n",
      " [ 14   1   0   7  24  83   7]\n",
      " [  5   2   5  13   6   8  13]]\n",
      "accuracy score =  0.298811544991511\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.19      0.30      0.23       127\n",
      "    disgust        0.05      0.08      0.06        24\n",
      "       fear        0.07      0.21      0.11        14\n",
      "        joy        0.34      0.36      0.35       255\n",
      "     others        0.50      0.21      0.30       570\n",
      "    sadness        0.31      0.61      0.41       136\n",
      "   surprise        0.11      0.25      0.15        52\n",
      "\n",
      "    accuracy                           0.30      1178\n",
      "   macro avg       0.23      0.29      0.23      1178\n",
      "weighted avg       0.38      0.30      0.30      1178\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x21c620b9340>"
      ]
     },
     "execution_count": 11,
     "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": [
    "#Naive Bayes\n",
    "#training model\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "nb = GaussianNB()\n",
    "nb.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = nb.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "nba = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(nb,x_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "caa8b806",
   "metadata": {},
   "source": [
    "## Árbol de Decisión y Bosques Aleatorios. \n",
    "Como modelos supervisados que son, a los árboles de decisión que clasifican los datos del conjunto de entrada en función de una variable clasificadora categórica (es decir, que toma un conjunto finito de valores) se los llama árboles de clasificación. Si la variable clasificadora es continua, hablaríamos de árboles de regresión. Ambos tipos de árboles de decisión dan el nombre común CART (Classification and Regression Trees) según el trabajo original descrito por Breiman et al \\cite{mt2}. Cuando los clasificadores base son árboles de decisión y se utiliza un muestreo tanto de los elementos del conjunto original de entrenamiento como de sus variables, el clasificador combinado se conoce como \\textit{Random forest} \\cite{mt3}, dado que se trata precisamente de un conjunto (o bosque) de árboles que han sido creados mediante un proceso aleatorio (por lo que respecta al conjunto de entrenamiento usado para cada uno de ellos). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "0e4f6879",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[ 31   4   1  17  57  12   5]\n",
      " [  7   4   0   2   7   3   1]\n",
      " [  1   0   3   0   5   0   5]\n",
      " [  8   2   0  96 127  11  11]\n",
      " [ 20   6   4  92 400  31  17]\n",
      " [ 11   1   0  10  52  62   0]\n",
      " [  2   2   1  19  23   1   4]]\n",
      "accuracy score =  0.5093378607809848\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.39      0.24      0.30       127\n",
      "    disgust        0.21      0.17      0.19        24\n",
      "       fear        0.33      0.21      0.26        14\n",
      "        joy        0.41      0.38      0.39       255\n",
      "     others        0.60      0.70      0.64       570\n",
      "    sadness        0.52      0.46      0.48       136\n",
      "   surprise        0.09      0.08      0.08        52\n",
      "\n",
      "    accuracy                           0.51      1178\n",
      "   macro avg       0.36      0.32      0.34      1178\n",
      "weighted avg       0.49      0.51      0.50      1178\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x20a0128d0d0>"
      ]
     },
     "execution_count": 11,
     "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": [
    "#Decision Tree\n",
    "#training model\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "dt = DecisionTreeClassifier(criterion = 'entropy')\n",
    "dt.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = dt.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "dta = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(dt,x_test,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "32da1c23",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[ 23   2   0   6  82  11   3]\n",
      " [  5   3   0   2  13   1   0]\n",
      " [  0   0   3   0  10   0   1]\n",
      " [  4   0   0  88 161   2   0]\n",
      " [  4   1   1  26 517  13   8]\n",
      " [  1   0   0   2  49  84   0]\n",
      " [  3   0   1   8  37   1   2]]\n",
      "accuracy score =  0.6112054329371817\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.57      0.18      0.28       127\n",
      "    disgust        0.50      0.12      0.20        24\n",
      "       fear        0.60      0.21      0.32        14\n",
      "        joy        0.67      0.35      0.45       255\n",
      "     others        0.59      0.91      0.72       570\n",
      "    sadness        0.75      0.62      0.68       136\n",
      "   surprise        0.14      0.04      0.06        52\n",
      "\n",
      "    accuracy                           0.61      1178\n",
      "   macro avg       0.55      0.35      0.39      1178\n",
      "weighted avg       0.60      0.61      0.56      1178\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x20a001fdaf0>"
      ]
     },
     "execution_count": 12,
     "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": [
    "#Random Forest\n",
    "#training model\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "rf = RandomForestClassifier(n_estimators = 140, criterion = 'entropy',random_state = 0)\n",
    "rf.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = rf.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "rfa = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(rf,x_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99ed1e2f",
   "metadata": {},
   "source": [
    "\n",
    "## CatBoost.\n",
    "Es el acronimo de Category  and Boosting. Este algoritmo puede mejorar el rendimiento del modelo al tiempo que reduce el sobreajuste y el tiempo dedicado al ajuste. CatBoost tiene varios parámetros para ajustar. Aún así, reduce la necesidad de un ajuste extensivo de hiperparámetros porque los parámetros predeterminados producen un gran resultado. Además, es una implementación novedosa de alto rendimiento y codicia que impulsa el gradiente."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "157debd9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate set to 0.085632\n",
      "0:\tlearn: 1.8416140\ttotal: 173ms\tremaining: 2m 52s\n",
      "1:\tlearn: 1.7591002\ttotal: 200ms\tremaining: 1m 39s\n",
      "2:\tlearn: 1.6948121\ttotal: 226ms\tremaining: 1m 15s\n",
      "3:\tlearn: 1.6433492\ttotal: 248ms\tremaining: 1m 1s\n",
      "4:\tlearn: 1.5999876\ttotal: 272ms\tremaining: 54.2s\n",
      "5:\tlearn: 1.5639437\ttotal: 294ms\tremaining: 48.8s\n",
      "6:\tlearn: 1.5318752\ttotal: 316ms\tremaining: 44.8s\n",
      "7:\tlearn: 1.5018282\ttotal: 338ms\tremaining: 41.9s\n",
      "8:\tlearn: 1.4741770\ttotal: 362ms\tremaining: 39.8s\n",
      "9:\tlearn: 1.4495541\ttotal: 383ms\tremaining: 38s\n",
      "10:\tlearn: 1.4294501\ttotal: 407ms\tremaining: 36.6s\n",
      "11:\tlearn: 1.4134752\ttotal: 435ms\tremaining: 35.8s\n",
      "12:\tlearn: 1.3981839\ttotal: 459ms\tremaining: 34.8s\n",
      "13:\tlearn: 1.3829343\ttotal: 481ms\tremaining: 33.9s\n",
      "14:\tlearn: 1.3675150\ttotal: 502ms\tremaining: 33s\n",
      "15:\tlearn: 1.3559730\ttotal: 525ms\tremaining: 32.3s\n",
      "16:\tlearn: 1.3464321\ttotal: 554ms\tremaining: 32s\n",
      "17:\tlearn: 1.3334058\ttotal: 580ms\tremaining: 31.6s\n",
      "18:\tlearn: 1.3234338\ttotal: 605ms\tremaining: 31.2s\n",
      "19:\tlearn: 1.3152995\ttotal: 628ms\tremaining: 30.8s\n",
      "20:\tlearn: 1.3085615\ttotal: 659ms\tremaining: 30.7s\n",
      "21:\tlearn: 1.3003797\ttotal: 806ms\tremaining: 35.8s\n",
      "22:\tlearn: 1.2943844\ttotal: 885ms\tremaining: 37.6s\n",
      "23:\tlearn: 1.2875508\ttotal: 948ms\tremaining: 38.6s\n",
      "24:\tlearn: 1.2822754\ttotal: 1.06s\tremaining: 41.5s\n",
      "25:\tlearn: 1.2772383\ttotal: 1.1s\tremaining: 41.2s\n",
      "26:\tlearn: 1.2730010\ttotal: 1.14s\tremaining: 41.1s\n",
      "27:\tlearn: 1.2691594\ttotal: 1.17s\tremaining: 40.8s\n",
      "28:\tlearn: 1.2645741\ttotal: 1.2s\tremaining: 40.3s\n",
      "29:\tlearn: 1.2607955\ttotal: 1.23s\tremaining: 39.8s\n",
      "30:\tlearn: 1.2558842\ttotal: 1.26s\tremaining: 39.4s\n",
      "31:\tlearn: 1.2521463\ttotal: 1.29s\tremaining: 38.9s\n",
      "32:\tlearn: 1.2485820\ttotal: 1.33s\tremaining: 38.9s\n",
      "33:\tlearn: 1.2449157\ttotal: 1.35s\tremaining: 38.5s\n",
      "34:\tlearn: 1.2426878\ttotal: 1.38s\tremaining: 38.1s\n",
      "35:\tlearn: 1.2398061\ttotal: 1.4s\tremaining: 37.5s\n",
      "36:\tlearn: 1.2370617\ttotal: 1.42s\tremaining: 37.1s\n",
      "37:\tlearn: 1.2338704\ttotal: 1.45s\tremaining: 36.7s\n",
      "38:\tlearn: 1.2296219\ttotal: 1.47s\tremaining: 36.2s\n",
      "39:\tlearn: 1.2270406\ttotal: 1.49s\tremaining: 35.8s\n",
      "40:\tlearn: 1.2252104\ttotal: 1.51s\tremaining: 35.4s\n",
      "41:\tlearn: 1.2226032\ttotal: 1.54s\tremaining: 35.1s\n",
      "42:\tlearn: 1.2199980\ttotal: 1.56s\tremaining: 34.8s\n",
      "43:\tlearn: 1.2171827\ttotal: 1.58s\tremaining: 34.4s\n",
      "44:\tlearn: 1.2138420\ttotal: 1.6s\tremaining: 34s\n",
      "45:\tlearn: 1.2125109\ttotal: 1.63s\tremaining: 33.7s\n",
      "46:\tlearn: 1.2106177\ttotal: 1.65s\tremaining: 33.5s\n",
      "47:\tlearn: 1.2084924\ttotal: 1.67s\tremaining: 33.2s\n",
      "48:\tlearn: 1.2062703\ttotal: 1.69s\tremaining: 32.9s\n",
      "49:\tlearn: 1.2047692\ttotal: 1.72s\tremaining: 32.6s\n",
      "50:\tlearn: 1.2022225\ttotal: 1.74s\tremaining: 32.4s\n",
      "51:\tlearn: 1.2007721\ttotal: 1.77s\tremaining: 32.2s\n",
      "52:\tlearn: 1.1992996\ttotal: 1.79s\tremaining: 32s\n",
      "53:\tlearn: 1.1976561\ttotal: 1.81s\tremaining: 31.7s\n",
      "54:\tlearn: 1.1958408\ttotal: 1.84s\tremaining: 31.6s\n",
      "55:\tlearn: 1.1946531\ttotal: 1.86s\tremaining: 31.4s\n",
      "56:\tlearn: 1.1922671\ttotal: 1.89s\tremaining: 31.2s\n",
      "57:\tlearn: 1.1912456\ttotal: 1.91s\tremaining: 31.1s\n",
      "58:\tlearn: 1.1891072\ttotal: 1.94s\tremaining: 30.9s\n",
      "59:\tlearn: 1.1867408\ttotal: 1.96s\tremaining: 30.8s\n",
      "60:\tlearn: 1.1845079\ttotal: 1.99s\tremaining: 30.6s\n",
      "61:\tlearn: 1.1839736\ttotal: 2.01s\tremaining: 30.4s\n",
      "62:\tlearn: 1.1832648\ttotal: 2.03s\tremaining: 30.2s\n",
      "63:\tlearn: 1.1814361\ttotal: 2.05s\tremaining: 30s\n",
      "64:\tlearn: 1.1799558\ttotal: 2.07s\tremaining: 29.8s\n",
      "65:\tlearn: 1.1786842\ttotal: 2.09s\tremaining: 29.6s\n",
      "66:\tlearn: 1.1771057\ttotal: 2.11s\tremaining: 29.4s\n",
      "67:\tlearn: 1.1761016\ttotal: 2.13s\tremaining: 29.2s\n",
      "68:\tlearn: 1.1744166\ttotal: 2.15s\tremaining: 29s\n",
      "69:\tlearn: 1.1733318\ttotal: 2.24s\tremaining: 29.7s\n",
      "70:\tlearn: 1.1719400\ttotal: 2.27s\tremaining: 29.7s\n",
      "71:\tlearn: 1.1710144\ttotal: 2.3s\tremaining: 29.6s\n",
      "72:\tlearn: 1.1702890\ttotal: 2.32s\tremaining: 29.5s\n",
      "73:\tlearn: 1.1695721\ttotal: 2.35s\tremaining: 29.4s\n",
      "74:\tlearn: 1.1679646\ttotal: 2.37s\tremaining: 29.2s\n",
      "75:\tlearn: 1.1663738\ttotal: 2.39s\tremaining: 29s\n",
      "76:\tlearn: 1.1648582\ttotal: 2.41s\tremaining: 28.9s\n",
      "77:\tlearn: 1.1641580\ttotal: 2.43s\tremaining: 28.7s\n",
      "78:\tlearn: 1.1632515\ttotal: 2.46s\tremaining: 28.6s\n",
      "79:\tlearn: 1.1619835\ttotal: 2.48s\tremaining: 28.5s\n",
      "80:\tlearn: 1.1606060\ttotal: 2.5s\tremaining: 28.3s\n",
      "81:\tlearn: 1.1594440\ttotal: 2.52s\tremaining: 28.2s\n",
      "82:\tlearn: 1.1586398\ttotal: 2.54s\tremaining: 28.1s\n",
      "83:\tlearn: 1.1577160\ttotal: 2.56s\tremaining: 27.9s\n",
      "84:\tlearn: 1.1566624\ttotal: 2.58s\tremaining: 27.8s\n",
      "85:\tlearn: 1.1552932\ttotal: 2.6s\tremaining: 27.7s\n",
      "86:\tlearn: 1.1540593\ttotal: 2.63s\tremaining: 27.6s\n",
      "87:\tlearn: 1.1524435\ttotal: 2.65s\tremaining: 27.5s\n",
      "88:\tlearn: 1.1511130\ttotal: 2.67s\tremaining: 27.4s\n",
      "89:\tlearn: 1.1501999\ttotal: 2.69s\tremaining: 27.2s\n",
      "90:\tlearn: 1.1492112\ttotal: 2.71s\tremaining: 27.1s\n",
      "91:\tlearn: 1.1484449\ttotal: 2.74s\tremaining: 27s\n",
      "92:\tlearn: 1.1468249\ttotal: 2.76s\tremaining: 26.9s\n",
      "93:\tlearn: 1.1458637\ttotal: 2.78s\tremaining: 26.8s\n",
      "94:\tlearn: 1.1441143\ttotal: 2.8s\tremaining: 26.7s\n",
      "95:\tlearn: 1.1423142\ttotal: 2.82s\tremaining: 26.6s\n",
      "96:\tlearn: 1.1407202\ttotal: 2.84s\tremaining: 26.4s\n",
      "97:\tlearn: 1.1395666\ttotal: 2.86s\tremaining: 26.3s\n",
      "98:\tlearn: 1.1378526\ttotal: 2.88s\tremaining: 26.3s\n",
      "99:\tlearn: 1.1364545\ttotal: 2.9s\tremaining: 26.1s\n",
      "100:\tlearn: 1.1351370\ttotal: 2.92s\tremaining: 26s\n",
      "101:\tlearn: 1.1339939\ttotal: 2.95s\tremaining: 25.9s\n",
      "102:\tlearn: 1.1328391\ttotal: 2.97s\tremaining: 25.8s\n",
      "103:\tlearn: 1.1313352\ttotal: 2.99s\tremaining: 25.8s\n",
      "104:\tlearn: 1.1302360\ttotal: 3.02s\tremaining: 25.7s\n",
      "105:\tlearn: 1.1282298\ttotal: 3.05s\tremaining: 25.7s\n",
      "106:\tlearn: 1.1268344\ttotal: 3.08s\tremaining: 25.7s\n",
      "107:\tlearn: 1.1259719\ttotal: 3.11s\tremaining: 25.7s\n",
      "108:\tlearn: 1.1247561\ttotal: 3.14s\tremaining: 25.7s\n",
      "109:\tlearn: 1.1237978\ttotal: 3.17s\tremaining: 25.7s\n",
      "110:\tlearn: 1.1224196\ttotal: 3.21s\tremaining: 25.7s\n",
      "111:\tlearn: 1.1208862\ttotal: 3.24s\tremaining: 25.7s\n",
      "112:\tlearn: 1.1198261\ttotal: 3.26s\tremaining: 25.6s\n",
      "113:\tlearn: 1.1189659\ttotal: 3.29s\tremaining: 25.5s\n",
      "114:\tlearn: 1.1183328\ttotal: 3.31s\tremaining: 25.5s\n",
      "115:\tlearn: 1.1173118\ttotal: 3.34s\tremaining: 25.4s\n",
      "116:\tlearn: 1.1162646\ttotal: 3.36s\tremaining: 25.4s\n",
      "117:\tlearn: 1.1150883\ttotal: 3.39s\tremaining: 25.3s\n",
      "118:\tlearn: 1.1141030\ttotal: 3.41s\tremaining: 25.2s\n",
      "119:\tlearn: 1.1129427\ttotal: 3.43s\tremaining: 25.2s\n",
      "120:\tlearn: 1.1120445\ttotal: 3.46s\tremaining: 25.1s\n",
      "121:\tlearn: 1.1111438\ttotal: 3.48s\tremaining: 25s\n",
      "122:\tlearn: 1.1095154\ttotal: 3.5s\tremaining: 25s\n",
      "123:\tlearn: 1.1085481\ttotal: 3.52s\tremaining: 24.9s\n",
      "124:\tlearn: 1.1074262\ttotal: 3.55s\tremaining: 24.9s\n",
      "125:\tlearn: 1.1063263\ttotal: 3.58s\tremaining: 24.8s\n",
      "126:\tlearn: 1.1048375\ttotal: 3.6s\tremaining: 24.8s\n",
      "127:\tlearn: 1.1036900\ttotal: 3.63s\tremaining: 24.7s\n",
      "128:\tlearn: 1.1028029\ttotal: 3.65s\tremaining: 24.6s\n",
      "129:\tlearn: 1.1017060\ttotal: 3.67s\tremaining: 24.6s\n",
      "130:\tlearn: 1.1008518\ttotal: 3.69s\tremaining: 24.5s\n",
      "131:\tlearn: 1.0998912\ttotal: 3.71s\tremaining: 24.4s\n",
      "132:\tlearn: 1.0989285\ttotal: 3.73s\tremaining: 24.3s\n",
      "133:\tlearn: 1.0977703\ttotal: 3.75s\tremaining: 24.3s\n",
      "134:\tlearn: 1.0964975\ttotal: 3.78s\tremaining: 24.2s\n",
      "135:\tlearn: 1.0956066\ttotal: 3.8s\tremaining: 24.1s\n",
      "136:\tlearn: 1.0945862\ttotal: 3.82s\tremaining: 24.1s\n",
      "137:\tlearn: 1.0935911\ttotal: 3.84s\tremaining: 24s\n",
      "138:\tlearn: 1.0924451\ttotal: 3.86s\tremaining: 23.9s\n",
      "139:\tlearn: 1.0916670\ttotal: 3.88s\tremaining: 23.8s\n",
      "140:\tlearn: 1.0907611\ttotal: 3.9s\tremaining: 23.8s\n",
      "141:\tlearn: 1.0896623\ttotal: 3.93s\tremaining: 23.8s\n",
      "142:\tlearn: 1.0885783\ttotal: 3.96s\tremaining: 23.7s\n",
      "143:\tlearn: 1.0876539\ttotal: 3.98s\tremaining: 23.7s\n",
      "144:\tlearn: 1.0864692\ttotal: 4.01s\tremaining: 23.6s\n",
      "145:\tlearn: 1.0853241\ttotal: 4.03s\tremaining: 23.6s\n",
      "146:\tlearn: 1.0843870\ttotal: 4.05s\tremaining: 23.5s\n",
      "147:\tlearn: 1.0836331\ttotal: 4.08s\tremaining: 23.5s\n",
      "148:\tlearn: 1.0828132\ttotal: 4.1s\tremaining: 23.4s\n",
      "149:\tlearn: 1.0821444\ttotal: 4.12s\tremaining: 23.4s\n",
      "150:\tlearn: 1.0811660\ttotal: 4.14s\tremaining: 23.3s\n",
      "151:\tlearn: 1.0806566\ttotal: 4.17s\tremaining: 23.2s\n",
      "152:\tlearn: 1.0797586\ttotal: 4.19s\tremaining: 23.2s\n",
      "153:\tlearn: 1.0787878\ttotal: 4.21s\tremaining: 23.1s\n",
      "154:\tlearn: 1.0781241\ttotal: 4.24s\tremaining: 23.1s\n",
      "155:\tlearn: 1.0771417\ttotal: 4.26s\tremaining: 23s\n",
      "156:\tlearn: 1.0763642\ttotal: 4.28s\tremaining: 23s\n",
      "157:\tlearn: 1.0753486\ttotal: 4.3s\tremaining: 22.9s\n",
      "158:\tlearn: 1.0747000\ttotal: 4.33s\tremaining: 22.9s\n",
      "159:\tlearn: 1.0738517\ttotal: 4.35s\tremaining: 22.8s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "160:\tlearn: 1.0727204\ttotal: 4.37s\tremaining: 22.8s\n",
      "161:\tlearn: 1.0716871\ttotal: 4.39s\tremaining: 22.7s\n",
      "162:\tlearn: 1.0709477\ttotal: 4.41s\tremaining: 22.6s\n",
      "163:\tlearn: 1.0703156\ttotal: 4.43s\tremaining: 22.6s\n",
      "164:\tlearn: 1.0695459\ttotal: 4.45s\tremaining: 22.5s\n",
      "165:\tlearn: 1.0690711\ttotal: 4.47s\tremaining: 22.5s\n",
      "166:\tlearn: 1.0678196\ttotal: 4.49s\tremaining: 22.4s\n",
      "167:\tlearn: 1.0666319\ttotal: 4.51s\tremaining: 22.3s\n",
      "168:\tlearn: 1.0661195\ttotal: 4.53s\tremaining: 22.3s\n",
      "169:\tlearn: 1.0653303\ttotal: 4.55s\tremaining: 22.2s\n",
      "170:\tlearn: 1.0646503\ttotal: 4.57s\tremaining: 22.2s\n",
      "171:\tlearn: 1.0639821\ttotal: 4.59s\tremaining: 22.1s\n",
      "172:\tlearn: 1.0634084\ttotal: 4.61s\tremaining: 22s\n",
      "173:\tlearn: 1.0625231\ttotal: 4.63s\tremaining: 22s\n",
      "174:\tlearn: 1.0617399\ttotal: 4.65s\tremaining: 21.9s\n",
      "175:\tlearn: 1.0609281\ttotal: 4.67s\tremaining: 21.9s\n",
      "176:\tlearn: 1.0603615\ttotal: 4.69s\tremaining: 21.8s\n",
      "177:\tlearn: 1.0598240\ttotal: 4.71s\tremaining: 21.8s\n",
      "178:\tlearn: 1.0591601\ttotal: 4.74s\tremaining: 21.7s\n",
      "179:\tlearn: 1.0585046\ttotal: 4.76s\tremaining: 21.7s\n",
      "180:\tlearn: 1.0578808\ttotal: 4.78s\tremaining: 21.6s\n",
      "181:\tlearn: 1.0573219\ttotal: 4.8s\tremaining: 21.6s\n",
      "182:\tlearn: 1.0566547\ttotal: 4.83s\tremaining: 21.5s\n",
      "183:\tlearn: 1.0558529\ttotal: 4.85s\tremaining: 21.5s\n",
      "184:\tlearn: 1.0553998\ttotal: 4.87s\tremaining: 21.5s\n",
      "185:\tlearn: 1.0547001\ttotal: 4.9s\tremaining: 21.5s\n",
      "186:\tlearn: 1.0540071\ttotal: 4.93s\tremaining: 21.5s\n",
      "187:\tlearn: 1.0532225\ttotal: 4.96s\tremaining: 21.4s\n",
      "188:\tlearn: 1.0523819\ttotal: 4.99s\tremaining: 21.4s\n",
      "189:\tlearn: 1.0519488\ttotal: 5.02s\tremaining: 21.4s\n",
      "190:\tlearn: 1.0512831\ttotal: 5.05s\tremaining: 21.4s\n",
      "191:\tlearn: 1.0503739\ttotal: 5.07s\tremaining: 21.3s\n",
      "192:\tlearn: 1.0500255\ttotal: 5.09s\tremaining: 21.3s\n",
      "193:\tlearn: 1.0496889\ttotal: 5.12s\tremaining: 21.3s\n",
      "194:\tlearn: 1.0492268\ttotal: 5.14s\tremaining: 21.2s\n",
      "195:\tlearn: 1.0487067\ttotal: 5.17s\tremaining: 21.2s\n",
      "196:\tlearn: 1.0479850\ttotal: 5.19s\tremaining: 21.2s\n",
      "197:\tlearn: 1.0473280\ttotal: 5.22s\tremaining: 21.2s\n",
      "198:\tlearn: 1.0469528\ttotal: 5.25s\tremaining: 21.1s\n",
      "199:\tlearn: 1.0465255\ttotal: 5.28s\tremaining: 21.1s\n",
      "200:\tlearn: 1.0458424\ttotal: 5.31s\tremaining: 21.1s\n",
      "201:\tlearn: 1.0454191\ttotal: 5.34s\tremaining: 21.1s\n",
      "202:\tlearn: 1.0448007\ttotal: 5.37s\tremaining: 21.1s\n",
      "203:\tlearn: 1.0443717\ttotal: 5.4s\tremaining: 21.1s\n",
      "204:\tlearn: 1.0439952\ttotal: 5.43s\tremaining: 21s\n",
      "205:\tlearn: 1.0430851\ttotal: 5.45s\tremaining: 21s\n",
      "206:\tlearn: 1.0426182\ttotal: 5.48s\tremaining: 21s\n",
      "207:\tlearn: 1.0421876\ttotal: 5.51s\tremaining: 21s\n",
      "208:\tlearn: 1.0419711\ttotal: 5.53s\tremaining: 20.9s\n",
      "209:\tlearn: 1.0412633\ttotal: 5.56s\tremaining: 20.9s\n",
      "210:\tlearn: 1.0403665\ttotal: 5.59s\tremaining: 20.9s\n",
      "211:\tlearn: 1.0397919\ttotal: 5.61s\tremaining: 20.9s\n",
      "212:\tlearn: 1.0392664\ttotal: 5.64s\tremaining: 20.8s\n",
      "213:\tlearn: 1.0387641\ttotal: 5.67s\tremaining: 20.8s\n",
      "214:\tlearn: 1.0379385\ttotal: 5.69s\tremaining: 20.8s\n",
      "215:\tlearn: 1.0369639\ttotal: 5.72s\tremaining: 20.8s\n",
      "216:\tlearn: 1.0365601\ttotal: 5.75s\tremaining: 20.7s\n",
      "217:\tlearn: 1.0361260\ttotal: 5.77s\tremaining: 20.7s\n",
      "218:\tlearn: 1.0356543\ttotal: 5.8s\tremaining: 20.7s\n",
      "219:\tlearn: 1.0353564\ttotal: 5.82s\tremaining: 20.6s\n",
      "220:\tlearn: 1.0346573\ttotal: 5.85s\tremaining: 20.6s\n",
      "221:\tlearn: 1.0339826\ttotal: 5.88s\tremaining: 20.6s\n",
      "222:\tlearn: 1.0332423\ttotal: 5.91s\tremaining: 20.6s\n",
      "223:\tlearn: 1.0327266\ttotal: 5.93s\tremaining: 20.5s\n",
      "224:\tlearn: 1.0323732\ttotal: 5.96s\tremaining: 20.5s\n",
      "225:\tlearn: 1.0318497\ttotal: 5.99s\tremaining: 20.5s\n",
      "226:\tlearn: 1.0313422\ttotal: 6.01s\tremaining: 20.5s\n",
      "227:\tlearn: 1.0308839\ttotal: 6.04s\tremaining: 20.4s\n",
      "228:\tlearn: 1.0303428\ttotal: 6.07s\tremaining: 20.4s\n",
      "229:\tlearn: 1.0297171\ttotal: 6.09s\tremaining: 20.4s\n",
      "230:\tlearn: 1.0293466\ttotal: 6.11s\tremaining: 20.3s\n",
      "231:\tlearn: 1.0288590\ttotal: 6.13s\tremaining: 20.3s\n",
      "232:\tlearn: 1.0280176\ttotal: 6.15s\tremaining: 20.2s\n",
      "233:\tlearn: 1.0274144\ttotal: 6.17s\tremaining: 20.2s\n",
      "234:\tlearn: 1.0272312\ttotal: 6.19s\tremaining: 20.1s\n",
      "235:\tlearn: 1.0268033\ttotal: 6.21s\tremaining: 20.1s\n",
      "236:\tlearn: 1.0260551\ttotal: 6.23s\tremaining: 20.1s\n",
      "237:\tlearn: 1.0253766\ttotal: 6.25s\tremaining: 20s\n",
      "238:\tlearn: 1.0248617\ttotal: 6.27s\tremaining: 20s\n",
      "239:\tlearn: 1.0245291\ttotal: 6.29s\tremaining: 19.9s\n",
      "240:\tlearn: 1.0239908\ttotal: 6.31s\tremaining: 19.9s\n",
      "241:\tlearn: 1.0234106\ttotal: 6.33s\tremaining: 19.8s\n",
      "242:\tlearn: 1.0228511\ttotal: 6.35s\tremaining: 19.8s\n",
      "243:\tlearn: 1.0224959\ttotal: 6.37s\tremaining: 19.7s\n",
      "244:\tlearn: 1.0221206\ttotal: 6.39s\tremaining: 19.7s\n",
      "245:\tlearn: 1.0216620\ttotal: 6.41s\tremaining: 19.6s\n",
      "246:\tlearn: 1.0211880\ttotal: 6.43s\tremaining: 19.6s\n",
      "247:\tlearn: 1.0207416\ttotal: 6.45s\tremaining: 19.5s\n",
      "248:\tlearn: 1.0199930\ttotal: 6.47s\tremaining: 19.5s\n",
      "249:\tlearn: 1.0195083\ttotal: 6.49s\tremaining: 19.5s\n",
      "250:\tlearn: 1.0186990\ttotal: 6.51s\tremaining: 19.4s\n",
      "251:\tlearn: 1.0183418\ttotal: 6.54s\tremaining: 19.4s\n",
      "252:\tlearn: 1.0179930\ttotal: 6.56s\tremaining: 19.4s\n",
      "253:\tlearn: 1.0173951\ttotal: 6.58s\tremaining: 19.3s\n",
      "254:\tlearn: 1.0166372\ttotal: 6.61s\tremaining: 19.3s\n",
      "255:\tlearn: 1.0161142\ttotal: 6.63s\tremaining: 19.3s\n",
      "256:\tlearn: 1.0158623\ttotal: 6.65s\tremaining: 19.2s\n",
      "257:\tlearn: 1.0154314\ttotal: 6.67s\tremaining: 19.2s\n",
      "258:\tlearn: 1.0150477\ttotal: 6.69s\tremaining: 19.1s\n",
      "259:\tlearn: 1.0148937\ttotal: 6.71s\tremaining: 19.1s\n",
      "260:\tlearn: 1.0144445\ttotal: 6.73s\tremaining: 19.1s\n",
      "261:\tlearn: 1.0140006\ttotal: 6.75s\tremaining: 19s\n",
      "262:\tlearn: 1.0135005\ttotal: 6.78s\tremaining: 19s\n",
      "263:\tlearn: 1.0131022\ttotal: 6.79s\tremaining: 18.9s\n",
      "264:\tlearn: 1.0126948\ttotal: 6.82s\tremaining: 18.9s\n",
      "265:\tlearn: 1.0121457\ttotal: 6.84s\tremaining: 18.9s\n",
      "266:\tlearn: 1.0116858\ttotal: 6.86s\tremaining: 18.8s\n",
      "267:\tlearn: 1.0113197\ttotal: 6.88s\tremaining: 18.8s\n",
      "268:\tlearn: 1.0110835\ttotal: 6.91s\tremaining: 18.8s\n",
      "269:\tlearn: 1.0105277\ttotal: 6.93s\tremaining: 18.7s\n",
      "270:\tlearn: 1.0100915\ttotal: 6.95s\tremaining: 18.7s\n",
      "271:\tlearn: 1.0094940\ttotal: 6.97s\tremaining: 18.7s\n",
      "272:\tlearn: 1.0092382\ttotal: 6.99s\tremaining: 18.6s\n",
      "273:\tlearn: 1.0087576\ttotal: 7.01s\tremaining: 18.6s\n",
      "274:\tlearn: 1.0081135\ttotal: 7.04s\tremaining: 18.6s\n",
      "275:\tlearn: 1.0077049\ttotal: 7.06s\tremaining: 18.5s\n",
      "276:\tlearn: 1.0073221\ttotal: 7.08s\tremaining: 18.5s\n",
      "277:\tlearn: 1.0068858\ttotal: 7.1s\tremaining: 18.4s\n",
      "278:\tlearn: 1.0061907\ttotal: 7.12s\tremaining: 18.4s\n",
      "279:\tlearn: 1.0058689\ttotal: 7.14s\tremaining: 18.4s\n",
      "280:\tlearn: 1.0056395\ttotal: 7.17s\tremaining: 18.3s\n",
      "281:\tlearn: 1.0052284\ttotal: 7.19s\tremaining: 18.3s\n",
      "282:\tlearn: 1.0050137\ttotal: 7.21s\tremaining: 18.3s\n",
      "283:\tlearn: 1.0045557\ttotal: 7.23s\tremaining: 18.2s\n",
      "284:\tlearn: 1.0039702\ttotal: 7.25s\tremaining: 18.2s\n",
      "285:\tlearn: 1.0037347\ttotal: 7.27s\tremaining: 18.2s\n",
      "286:\tlearn: 1.0031822\ttotal: 7.29s\tremaining: 18.1s\n",
      "287:\tlearn: 1.0026634\ttotal: 7.31s\tremaining: 18.1s\n",
      "288:\tlearn: 1.0023253\ttotal: 7.33s\tremaining: 18s\n",
      "289:\tlearn: 1.0018778\ttotal: 7.35s\tremaining: 18s\n",
      "290:\tlearn: 1.0015846\ttotal: 7.37s\tremaining: 18s\n",
      "291:\tlearn: 1.0010242\ttotal: 7.39s\tremaining: 17.9s\n",
      "292:\tlearn: 1.0006218\ttotal: 7.41s\tremaining: 17.9s\n",
      "293:\tlearn: 1.0000740\ttotal: 7.43s\tremaining: 17.9s\n",
      "294:\tlearn: 0.9996639\ttotal: 7.45s\tremaining: 17.8s\n",
      "295:\tlearn: 0.9992627\ttotal: 7.47s\tremaining: 17.8s\n",
      "296:\tlearn: 0.9987715\ttotal: 7.49s\tremaining: 17.7s\n",
      "297:\tlearn: 0.9980593\ttotal: 7.51s\tremaining: 17.7s\n",
      "298:\tlearn: 0.9975103\ttotal: 7.53s\tremaining: 17.7s\n",
      "299:\tlearn: 0.9970284\ttotal: 7.55s\tremaining: 17.6s\n",
      "300:\tlearn: 0.9966871\ttotal: 7.57s\tremaining: 17.6s\n",
      "301:\tlearn: 0.9960179\ttotal: 7.59s\tremaining: 17.5s\n",
      "302:\tlearn: 0.9956699\ttotal: 7.61s\tremaining: 17.5s\n",
      "303:\tlearn: 0.9953399\ttotal: 7.63s\tremaining: 17.5s\n",
      "304:\tlearn: 0.9948711\ttotal: 7.65s\tremaining: 17.4s\n",
      "305:\tlearn: 0.9944334\ttotal: 7.67s\tremaining: 17.4s\n",
      "306:\tlearn: 0.9939736\ttotal: 7.69s\tremaining: 17.4s\n",
      "307:\tlearn: 0.9934686\ttotal: 7.72s\tremaining: 17.3s\n",
      "308:\tlearn: 0.9930535\ttotal: 7.74s\tremaining: 17.3s\n",
      "309:\tlearn: 0.9924838\ttotal: 7.76s\tremaining: 17.3s\n",
      "310:\tlearn: 0.9921080\ttotal: 7.78s\tremaining: 17.2s\n",
      "311:\tlearn: 0.9915963\ttotal: 7.8s\tremaining: 17.2s\n",
      "312:\tlearn: 0.9912714\ttotal: 7.82s\tremaining: 17.2s\n",
      "313:\tlearn: 0.9909431\ttotal: 7.84s\tremaining: 17.1s\n",
      "314:\tlearn: 0.9905953\ttotal: 7.86s\tremaining: 17.1s\n",
      "315:\tlearn: 0.9900996\ttotal: 7.88s\tremaining: 17.1s\n",
      "316:\tlearn: 0.9897974\ttotal: 7.9s\tremaining: 17s\n",
      "317:\tlearn: 0.9892317\ttotal: 7.92s\tremaining: 17s\n",
      "318:\tlearn: 0.9888715\ttotal: 7.95s\tremaining: 17s\n",
      "319:\tlearn: 0.9885944\ttotal: 7.96s\tremaining: 16.9s\n",
      "320:\tlearn: 0.9881760\ttotal: 7.99s\tremaining: 16.9s\n",
      "321:\tlearn: 0.9879839\ttotal: 8.01s\tremaining: 16.9s\n",
      "322:\tlearn: 0.9875173\ttotal: 8.03s\tremaining: 16.8s\n",
      "323:\tlearn: 0.9870740\ttotal: 8.05s\tremaining: 16.8s\n",
      "324:\tlearn: 0.9867659\ttotal: 8.07s\tremaining: 16.8s\n",
      "325:\tlearn: 0.9862904\ttotal: 8.09s\tremaining: 16.7s\n",
      "326:\tlearn: 0.9860166\ttotal: 8.11s\tremaining: 16.7s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "327:\tlearn: 0.9856165\ttotal: 8.13s\tremaining: 16.7s\n",
      "328:\tlearn: 0.9851034\ttotal: 8.16s\tremaining: 16.6s\n",
      "329:\tlearn: 0.9848004\ttotal: 8.18s\tremaining: 16.6s\n",
      "330:\tlearn: 0.9845022\ttotal: 8.21s\tremaining: 16.6s\n",
      "331:\tlearn: 0.9842536\ttotal: 8.23s\tremaining: 16.6s\n",
      "332:\tlearn: 0.9838841\ttotal: 8.25s\tremaining: 16.5s\n",
      "333:\tlearn: 0.9836543\ttotal: 8.27s\tremaining: 16.5s\n",
      "334:\tlearn: 0.9834125\ttotal: 8.3s\tremaining: 16.5s\n",
      "335:\tlearn: 0.9831116\ttotal: 8.32s\tremaining: 16.4s\n",
      "336:\tlearn: 0.9823862\ttotal: 8.34s\tremaining: 16.4s\n",
      "337:\tlearn: 0.9820862\ttotal: 8.36s\tremaining: 16.4s\n",
      "338:\tlearn: 0.9817640\ttotal: 8.38s\tremaining: 16.3s\n",
      "339:\tlearn: 0.9813659\ttotal: 8.4s\tremaining: 16.3s\n",
      "340:\tlearn: 0.9811366\ttotal: 8.42s\tremaining: 16.3s\n",
      "341:\tlearn: 0.9806632\ttotal: 8.45s\tremaining: 16.2s\n",
      "342:\tlearn: 0.9801268\ttotal: 8.47s\tremaining: 16.2s\n",
      "343:\tlearn: 0.9797956\ttotal: 8.49s\tremaining: 16.2s\n",
      "344:\tlearn: 0.9795316\ttotal: 8.51s\tremaining: 16.2s\n",
      "345:\tlearn: 0.9790765\ttotal: 8.53s\tremaining: 16.1s\n",
      "346:\tlearn: 0.9787797\ttotal: 8.56s\tremaining: 16.1s\n",
      "347:\tlearn: 0.9783137\ttotal: 8.58s\tremaining: 16.1s\n",
      "348:\tlearn: 0.9779228\ttotal: 8.61s\tremaining: 16.1s\n",
      "349:\tlearn: 0.9774218\ttotal: 8.63s\tremaining: 16s\n",
      "350:\tlearn: 0.9772182\ttotal: 8.65s\tremaining: 16s\n",
      "351:\tlearn: 0.9771124\ttotal: 8.67s\tremaining: 16s\n",
      "352:\tlearn: 0.9768446\ttotal: 8.7s\tremaining: 15.9s\n",
      "353:\tlearn: 0.9764352\ttotal: 8.72s\tremaining: 15.9s\n",
      "354:\tlearn: 0.9760465\ttotal: 8.74s\tremaining: 15.9s\n",
      "355:\tlearn: 0.9757038\ttotal: 8.77s\tremaining: 15.9s\n",
      "356:\tlearn: 0.9753917\ttotal: 8.79s\tremaining: 15.8s\n",
      "357:\tlearn: 0.9750099\ttotal: 8.81s\tremaining: 15.8s\n",
      "358:\tlearn: 0.9746937\ttotal: 8.83s\tremaining: 15.8s\n",
      "359:\tlearn: 0.9744282\ttotal: 8.85s\tremaining: 15.7s\n",
      "360:\tlearn: 0.9738963\ttotal: 8.88s\tremaining: 15.7s\n",
      "361:\tlearn: 0.9732718\ttotal: 8.9s\tremaining: 15.7s\n",
      "362:\tlearn: 0.9729830\ttotal: 8.92s\tremaining: 15.6s\n",
      "363:\tlearn: 0.9726942\ttotal: 8.94s\tremaining: 15.6s\n",
      "364:\tlearn: 0.9723106\ttotal: 8.96s\tremaining: 15.6s\n",
      "365:\tlearn: 0.9720202\ttotal: 8.98s\tremaining: 15.6s\n",
      "366:\tlearn: 0.9715288\ttotal: 9s\tremaining: 15.5s\n",
      "367:\tlearn: 0.9712002\ttotal: 9.02s\tremaining: 15.5s\n",
      "368:\tlearn: 0.9707461\ttotal: 9.04s\tremaining: 15.5s\n",
      "369:\tlearn: 0.9704002\ttotal: 9.06s\tremaining: 15.4s\n",
      "370:\tlearn: 0.9698374\ttotal: 9.08s\tremaining: 15.4s\n",
      "371:\tlearn: 0.9695145\ttotal: 9.11s\tremaining: 15.4s\n",
      "372:\tlearn: 0.9690846\ttotal: 9.13s\tremaining: 15.3s\n",
      "373:\tlearn: 0.9687592\ttotal: 9.15s\tremaining: 15.3s\n",
      "374:\tlearn: 0.9684491\ttotal: 9.17s\tremaining: 15.3s\n",
      "375:\tlearn: 0.9682268\ttotal: 9.19s\tremaining: 15.2s\n",
      "376:\tlearn: 0.9679099\ttotal: 9.21s\tremaining: 15.2s\n",
      "377:\tlearn: 0.9675946\ttotal: 9.23s\tremaining: 15.2s\n",
      "378:\tlearn: 0.9673389\ttotal: 9.25s\tremaining: 15.2s\n",
      "379:\tlearn: 0.9669032\ttotal: 9.27s\tremaining: 15.1s\n",
      "380:\tlearn: 0.9665585\ttotal: 9.29s\tremaining: 15.1s\n",
      "381:\tlearn: 0.9662475\ttotal: 9.31s\tremaining: 15.1s\n",
      "382:\tlearn: 0.9659225\ttotal: 9.33s\tremaining: 15s\n",
      "383:\tlearn: 0.9656730\ttotal: 9.35s\tremaining: 15s\n",
      "384:\tlearn: 0.9651743\ttotal: 9.37s\tremaining: 15s\n",
      "385:\tlearn: 0.9648692\ttotal: 9.39s\tremaining: 14.9s\n",
      "386:\tlearn: 0.9644289\ttotal: 9.41s\tremaining: 14.9s\n",
      "387:\tlearn: 0.9641979\ttotal: 9.44s\tremaining: 14.9s\n",
      "388:\tlearn: 0.9639152\ttotal: 9.46s\tremaining: 14.9s\n",
      "389:\tlearn: 0.9636038\ttotal: 9.47s\tremaining: 14.8s\n",
      "390:\tlearn: 0.9631108\ttotal: 9.49s\tremaining: 14.8s\n",
      "391:\tlearn: 0.9626753\ttotal: 9.52s\tremaining: 14.8s\n",
      "392:\tlearn: 0.9622925\ttotal: 9.53s\tremaining: 14.7s\n",
      "393:\tlearn: 0.9621852\ttotal: 9.55s\tremaining: 14.7s\n",
      "394:\tlearn: 0.9616915\ttotal: 9.57s\tremaining: 14.7s\n",
      "395:\tlearn: 0.9612166\ttotal: 9.6s\tremaining: 14.6s\n",
      "396:\tlearn: 0.9608114\ttotal: 9.62s\tremaining: 14.6s\n",
      "397:\tlearn: 0.9603231\ttotal: 9.64s\tremaining: 14.6s\n",
      "398:\tlearn: 0.9600178\ttotal: 9.66s\tremaining: 14.5s\n",
      "399:\tlearn: 0.9597621\ttotal: 9.68s\tremaining: 14.5s\n",
      "400:\tlearn: 0.9594236\ttotal: 9.7s\tremaining: 14.5s\n",
      "401:\tlearn: 0.9591246\ttotal: 9.71s\tremaining: 14.4s\n",
      "402:\tlearn: 0.9588713\ttotal: 9.73s\tremaining: 14.4s\n",
      "403:\tlearn: 0.9584781\ttotal: 9.75s\tremaining: 14.4s\n",
      "404:\tlearn: 0.9582072\ttotal: 9.77s\tremaining: 14.4s\n",
      "405:\tlearn: 0.9577337\ttotal: 9.79s\tremaining: 14.3s\n",
      "406:\tlearn: 0.9574354\ttotal: 9.81s\tremaining: 14.3s\n",
      "407:\tlearn: 0.9571751\ttotal: 9.83s\tremaining: 14.3s\n",
      "408:\tlearn: 0.9570133\ttotal: 9.85s\tremaining: 14.2s\n",
      "409:\tlearn: 0.9564735\ttotal: 9.87s\tremaining: 14.2s\n",
      "410:\tlearn: 0.9562363\ttotal: 9.9s\tremaining: 14.2s\n",
      "411:\tlearn: 0.9558968\ttotal: 9.92s\tremaining: 14.2s\n",
      "412:\tlearn: 0.9555761\ttotal: 9.95s\tremaining: 14.1s\n",
      "413:\tlearn: 0.9552887\ttotal: 9.97s\tremaining: 14.1s\n",
      "414:\tlearn: 0.9551203\ttotal: 9.99s\tremaining: 14.1s\n",
      "415:\tlearn: 0.9547882\ttotal: 10s\tremaining: 14.1s\n",
      "416:\tlearn: 0.9545842\ttotal: 10s\tremaining: 14s\n",
      "417:\tlearn: 0.9543401\ttotal: 10.1s\tremaining: 14s\n",
      "418:\tlearn: 0.9540798\ttotal: 10.1s\tremaining: 14s\n",
      "419:\tlearn: 0.9537078\ttotal: 10.1s\tremaining: 14s\n",
      "420:\tlearn: 0.9533691\ttotal: 10.1s\tremaining: 13.9s\n",
      "421:\tlearn: 0.9531766\ttotal: 10.1s\tremaining: 13.9s\n",
      "422:\tlearn: 0.9529428\ttotal: 10.2s\tremaining: 13.9s\n",
      "423:\tlearn: 0.9526137\ttotal: 10.2s\tremaining: 13.9s\n",
      "424:\tlearn: 0.9523596\ttotal: 10.2s\tremaining: 13.8s\n",
      "425:\tlearn: 0.9519738\ttotal: 10.3s\tremaining: 13.8s\n",
      "426:\tlearn: 0.9514595\ttotal: 10.3s\tremaining: 13.8s\n",
      "427:\tlearn: 0.9511938\ttotal: 10.3s\tremaining: 13.8s\n",
      "428:\tlearn: 0.9508889\ttotal: 10.3s\tremaining: 13.7s\n",
      "429:\tlearn: 0.9504063\ttotal: 10.3s\tremaining: 13.7s\n",
      "430:\tlearn: 0.9498037\ttotal: 10.4s\tremaining: 13.7s\n",
      "431:\tlearn: 0.9495297\ttotal: 10.4s\tremaining: 13.7s\n",
      "432:\tlearn: 0.9492824\ttotal: 10.5s\tremaining: 13.7s\n",
      "433:\tlearn: 0.9489868\ttotal: 10.5s\tremaining: 13.7s\n",
      "434:\tlearn: 0.9485924\ttotal: 10.5s\tremaining: 13.6s\n",
      "435:\tlearn: 0.9480669\ttotal: 10.5s\tremaining: 13.6s\n",
      "436:\tlearn: 0.9476310\ttotal: 10.5s\tremaining: 13.6s\n",
      "437:\tlearn: 0.9474176\ttotal: 10.6s\tremaining: 13.6s\n",
      "438:\tlearn: 0.9471874\ttotal: 10.6s\tremaining: 13.5s\n",
      "439:\tlearn: 0.9469727\ttotal: 10.6s\tremaining: 13.5s\n",
      "440:\tlearn: 0.9467245\ttotal: 10.6s\tremaining: 13.5s\n",
      "441:\tlearn: 0.9463926\ttotal: 10.6s\tremaining: 13.4s\n",
      "442:\tlearn: 0.9460540\ttotal: 10.7s\tremaining: 13.4s\n",
      "443:\tlearn: 0.9458033\ttotal: 10.7s\tremaining: 13.4s\n",
      "444:\tlearn: 0.9456950\ttotal: 10.7s\tremaining: 13.4s\n",
      "445:\tlearn: 0.9453814\ttotal: 10.7s\tremaining: 13.3s\n",
      "446:\tlearn: 0.9450830\ttotal: 10.8s\tremaining: 13.3s\n",
      "447:\tlearn: 0.9445008\ttotal: 10.8s\tremaining: 13.3s\n",
      "448:\tlearn: 0.9442817\ttotal: 10.8s\tremaining: 13.2s\n",
      "449:\tlearn: 0.9439908\ttotal: 10.8s\tremaining: 13.2s\n",
      "450:\tlearn: 0.9437494\ttotal: 10.8s\tremaining: 13.2s\n",
      "451:\tlearn: 0.9433778\ttotal: 10.9s\tremaining: 13.2s\n",
      "452:\tlearn: 0.9431402\ttotal: 10.9s\tremaining: 13.1s\n",
      "453:\tlearn: 0.9429156\ttotal: 10.9s\tremaining: 13.1s\n",
      "454:\tlearn: 0.9426199\ttotal: 10.9s\tremaining: 13.1s\n",
      "455:\tlearn: 0.9422287\ttotal: 11s\tremaining: 13.1s\n",
      "456:\tlearn: 0.9419144\ttotal: 11s\tremaining: 13s\n",
      "457:\tlearn: 0.9416196\ttotal: 11s\tremaining: 13s\n",
      "458:\tlearn: 0.9414012\ttotal: 11s\tremaining: 13s\n",
      "459:\tlearn: 0.9409923\ttotal: 11.1s\tremaining: 13s\n",
      "460:\tlearn: 0.9408217\ttotal: 11.1s\tremaining: 13s\n",
      "461:\tlearn: 0.9404362\ttotal: 11.1s\tremaining: 12.9s\n",
      "462:\tlearn: 0.9402492\ttotal: 11.1s\tremaining: 12.9s\n",
      "463:\tlearn: 0.9401200\ttotal: 11.1s\tremaining: 12.9s\n",
      "464:\tlearn: 0.9398364\ttotal: 11.2s\tremaining: 12.8s\n",
      "465:\tlearn: 0.9395469\ttotal: 11.2s\tremaining: 12.8s\n",
      "466:\tlearn: 0.9390363\ttotal: 11.2s\tremaining: 12.8s\n",
      "467:\tlearn: 0.9386079\ttotal: 11.2s\tremaining: 12.8s\n",
      "468:\tlearn: 0.9382591\ttotal: 11.2s\tremaining: 12.7s\n",
      "469:\tlearn: 0.9379719\ttotal: 11.3s\tremaining: 12.7s\n",
      "470:\tlearn: 0.9378550\ttotal: 11.3s\tremaining: 12.7s\n",
      "471:\tlearn: 0.9376245\ttotal: 11.3s\tremaining: 12.6s\n",
      "472:\tlearn: 0.9372027\ttotal: 11.3s\tremaining: 12.6s\n",
      "473:\tlearn: 0.9370150\ttotal: 11.3s\tremaining: 12.6s\n",
      "474:\tlearn: 0.9367383\ttotal: 11.4s\tremaining: 12.6s\n",
      "475:\tlearn: 0.9364767\ttotal: 11.4s\tremaining: 12.5s\n",
      "476:\tlearn: 0.9362446\ttotal: 11.4s\tremaining: 12.5s\n",
      "477:\tlearn: 0.9360550\ttotal: 11.4s\tremaining: 12.5s\n",
      "478:\tlearn: 0.9355677\ttotal: 11.5s\tremaining: 12.5s\n",
      "479:\tlearn: 0.9352165\ttotal: 11.5s\tremaining: 12.5s\n",
      "480:\tlearn: 0.9349998\ttotal: 11.5s\tremaining: 12.4s\n",
      "481:\tlearn: 0.9347211\ttotal: 11.5s\tremaining: 12.4s\n",
      "482:\tlearn: 0.9345857\ttotal: 11.6s\tremaining: 12.4s\n",
      "483:\tlearn: 0.9343317\ttotal: 11.6s\tremaining: 12.3s\n",
      "484:\tlearn: 0.9340537\ttotal: 11.6s\tremaining: 12.3s\n",
      "485:\tlearn: 0.9337500\ttotal: 11.6s\tremaining: 12.3s\n",
      "486:\tlearn: 0.9335659\ttotal: 11.7s\tremaining: 12.3s\n",
      "487:\tlearn: 0.9332516\ttotal: 11.7s\tremaining: 12.2s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "488:\tlearn: 0.9330587\ttotal: 11.7s\tremaining: 12.2s\n",
      "489:\tlearn: 0.9328125\ttotal: 11.7s\tremaining: 12.2s\n",
      "490:\tlearn: 0.9324843\ttotal: 11.7s\tremaining: 12.2s\n",
      "491:\tlearn: 0.9323194\ttotal: 11.8s\tremaining: 12.1s\n",
      "492:\tlearn: 0.9320305\ttotal: 11.8s\tremaining: 12.1s\n",
      "493:\tlearn: 0.9316925\ttotal: 11.8s\tremaining: 12.1s\n",
      "494:\tlearn: 0.9312431\ttotal: 11.8s\tremaining: 12.1s\n",
      "495:\tlearn: 0.9311077\ttotal: 11.8s\tremaining: 12s\n",
      "496:\tlearn: 0.9308265\ttotal: 11.9s\tremaining: 12s\n",
      "497:\tlearn: 0.9306159\ttotal: 11.9s\tremaining: 12s\n",
      "498:\tlearn: 0.9303941\ttotal: 11.9s\tremaining: 12s\n",
      "499:\tlearn: 0.9302361\ttotal: 11.9s\tremaining: 11.9s\n",
      "500:\tlearn: 0.9300395\ttotal: 12s\tremaining: 11.9s\n",
      "501:\tlearn: 0.9297907\ttotal: 12s\tremaining: 11.9s\n",
      "502:\tlearn: 0.9292787\ttotal: 12s\tremaining: 11.9s\n",
      "503:\tlearn: 0.9290564\ttotal: 12s\tremaining: 11.8s\n",
      "504:\tlearn: 0.9287123\ttotal: 12s\tremaining: 11.8s\n",
      "505:\tlearn: 0.9283232\ttotal: 12.1s\tremaining: 11.8s\n",
      "506:\tlearn: 0.9278807\ttotal: 12.1s\tremaining: 11.8s\n",
      "507:\tlearn: 0.9275096\ttotal: 12.1s\tremaining: 11.7s\n",
      "508:\tlearn: 0.9272625\ttotal: 12.1s\tremaining: 11.7s\n",
      "509:\tlearn: 0.9269935\ttotal: 12.1s\tremaining: 11.7s\n",
      "510:\tlearn: 0.9267711\ttotal: 12.2s\tremaining: 11.6s\n",
      "511:\tlearn: 0.9264010\ttotal: 12.2s\tremaining: 11.6s\n",
      "512:\tlearn: 0.9261280\ttotal: 12.2s\tremaining: 11.6s\n",
      "513:\tlearn: 0.9258624\ttotal: 12.2s\tremaining: 11.6s\n",
      "514:\tlearn: 0.9255030\ttotal: 12.3s\tremaining: 11.5s\n",
      "515:\tlearn: 0.9252389\ttotal: 12.3s\tremaining: 11.5s\n",
      "516:\tlearn: 0.9249911\ttotal: 12.3s\tremaining: 11.5s\n",
      "517:\tlearn: 0.9246918\ttotal: 12.3s\tremaining: 11.5s\n",
      "518:\tlearn: 0.9244724\ttotal: 12.3s\tremaining: 11.4s\n",
      "519:\tlearn: 0.9242802\ttotal: 12.4s\tremaining: 11.4s\n",
      "520:\tlearn: 0.9239161\ttotal: 12.4s\tremaining: 11.4s\n",
      "521:\tlearn: 0.9236676\ttotal: 12.4s\tremaining: 11.4s\n",
      "522:\tlearn: 0.9234223\ttotal: 12.4s\tremaining: 11.3s\n",
      "523:\tlearn: 0.9231798\ttotal: 12.4s\tremaining: 11.3s\n",
      "524:\tlearn: 0.9230607\ttotal: 12.5s\tremaining: 11.3s\n",
      "525:\tlearn: 0.9229469\ttotal: 12.5s\tremaining: 11.2s\n",
      "526:\tlearn: 0.9227502\ttotal: 12.5s\tremaining: 11.2s\n",
      "527:\tlearn: 0.9226068\ttotal: 12.5s\tremaining: 11.2s\n",
      "528:\tlearn: 0.9222615\ttotal: 12.5s\tremaining: 11.2s\n",
      "529:\tlearn: 0.9221588\ttotal: 12.6s\tremaining: 11.1s\n",
      "530:\tlearn: 0.9218696\ttotal: 12.6s\tremaining: 11.1s\n",
      "531:\tlearn: 0.9212916\ttotal: 12.6s\tremaining: 11.1s\n",
      "532:\tlearn: 0.9209256\ttotal: 12.6s\tremaining: 11.1s\n",
      "533:\tlearn: 0.9207449\ttotal: 12.6s\tremaining: 11s\n",
      "534:\tlearn: 0.9205925\ttotal: 12.7s\tremaining: 11s\n",
      "535:\tlearn: 0.9203923\ttotal: 12.7s\tremaining: 11s\n",
      "536:\tlearn: 0.9201417\ttotal: 12.7s\tremaining: 10.9s\n",
      "537:\tlearn: 0.9198220\ttotal: 12.7s\tremaining: 10.9s\n",
      "538:\tlearn: 0.9196317\ttotal: 12.7s\tremaining: 10.9s\n",
      "539:\tlearn: 0.9193594\ttotal: 12.8s\tremaining: 10.9s\n",
      "540:\tlearn: 0.9192120\ttotal: 12.8s\tremaining: 10.8s\n",
      "541:\tlearn: 0.9189099\ttotal: 12.8s\tremaining: 10.8s\n",
      "542:\tlearn: 0.9187838\ttotal: 12.8s\tremaining: 10.8s\n",
      "543:\tlearn: 0.9184093\ttotal: 12.8s\tremaining: 10.8s\n",
      "544:\tlearn: 0.9181335\ttotal: 12.8s\tremaining: 10.7s\n",
      "545:\tlearn: 0.9179252\ttotal: 12.9s\tremaining: 10.7s\n",
      "546:\tlearn: 0.9177176\ttotal: 12.9s\tremaining: 10.7s\n",
      "547:\tlearn: 0.9174479\ttotal: 12.9s\tremaining: 10.6s\n",
      "548:\tlearn: 0.9171638\ttotal: 12.9s\tremaining: 10.6s\n",
      "549:\tlearn: 0.9169435\ttotal: 12.9s\tremaining: 10.6s\n",
      "550:\tlearn: 0.9166963\ttotal: 13s\tremaining: 10.6s\n",
      "551:\tlearn: 0.9163801\ttotal: 13s\tremaining: 10.5s\n",
      "552:\tlearn: 0.9161816\ttotal: 13s\tremaining: 10.5s\n",
      "553:\tlearn: 0.9159360\ttotal: 13s\tremaining: 10.5s\n",
      "554:\tlearn: 0.9156293\ttotal: 13s\tremaining: 10.5s\n",
      "555:\tlearn: 0.9152819\ttotal: 13.1s\tremaining: 10.4s\n",
      "556:\tlearn: 0.9149370\ttotal: 13.1s\tremaining: 10.4s\n",
      "557:\tlearn: 0.9146731\ttotal: 13.1s\tremaining: 10.4s\n",
      "558:\tlearn: 0.9144481\ttotal: 13.1s\tremaining: 10.4s\n",
      "559:\tlearn: 0.9143334\ttotal: 13.1s\tremaining: 10.3s\n",
      "560:\tlearn: 0.9139117\ttotal: 13.2s\tremaining: 10.3s\n",
      "561:\tlearn: 0.9136907\ttotal: 13.2s\tremaining: 10.3s\n",
      "562:\tlearn: 0.9133089\ttotal: 13.2s\tremaining: 10.2s\n",
      "563:\tlearn: 0.9131446\ttotal: 13.2s\tremaining: 10.2s\n",
      "564:\tlearn: 0.9128468\ttotal: 13.2s\tremaining: 10.2s\n",
      "565:\tlearn: 0.9126891\ttotal: 13.3s\tremaining: 10.2s\n",
      "566:\tlearn: 0.9124467\ttotal: 13.3s\tremaining: 10.1s\n",
      "567:\tlearn: 0.9122465\ttotal: 13.3s\tremaining: 10.1s\n",
      "568:\tlearn: 0.9120169\ttotal: 13.3s\tremaining: 10.1s\n",
      "569:\tlearn: 0.9116788\ttotal: 13.4s\tremaining: 10.1s\n",
      "570:\tlearn: 0.9114944\ttotal: 13.4s\tremaining: 10s\n",
      "571:\tlearn: 0.9113988\ttotal: 13.4s\tremaining: 10s\n",
      "572:\tlearn: 0.9112259\ttotal: 13.4s\tremaining: 10s\n",
      "573:\tlearn: 0.9111380\ttotal: 13.4s\tremaining: 9.97s\n",
      "574:\tlearn: 0.9110054\ttotal: 13.5s\tremaining: 9.95s\n",
      "575:\tlearn: 0.9107021\ttotal: 13.5s\tremaining: 9.93s\n",
      "576:\tlearn: 0.9104435\ttotal: 13.5s\tremaining: 9.9s\n",
      "577:\tlearn: 0.9101672\ttotal: 13.5s\tremaining: 9.88s\n",
      "578:\tlearn: 0.9100374\ttotal: 13.5s\tremaining: 9.85s\n",
      "579:\tlearn: 0.9096668\ttotal: 13.6s\tremaining: 9.82s\n",
      "580:\tlearn: 0.9094737\ttotal: 13.6s\tremaining: 9.8s\n",
      "581:\tlearn: 0.9092286\ttotal: 13.6s\tremaining: 9.78s\n",
      "582:\tlearn: 0.9090182\ttotal: 13.6s\tremaining: 9.75s\n",
      "583:\tlearn: 0.9087845\ttotal: 13.7s\tremaining: 9.74s\n",
      "584:\tlearn: 0.9085571\ttotal: 13.7s\tremaining: 9.72s\n",
      "585:\tlearn: 0.9081656\ttotal: 13.7s\tremaining: 9.69s\n",
      "586:\tlearn: 0.9078209\ttotal: 13.7s\tremaining: 9.67s\n",
      "587:\tlearn: 0.9075218\ttotal: 13.8s\tremaining: 9.64s\n",
      "588:\tlearn: 0.9072039\ttotal: 13.8s\tremaining: 9.62s\n",
      "589:\tlearn: 0.9069787\ttotal: 13.8s\tremaining: 9.59s\n",
      "590:\tlearn: 0.9066278\ttotal: 13.8s\tremaining: 9.57s\n",
      "591:\tlearn: 0.9064366\ttotal: 13.8s\tremaining: 9.54s\n",
      "592:\tlearn: 0.9062862\ttotal: 13.9s\tremaining: 9.52s\n",
      "593:\tlearn: 0.9060634\ttotal: 13.9s\tremaining: 9.5s\n",
      "594:\tlearn: 0.9059091\ttotal: 13.9s\tremaining: 9.47s\n",
      "595:\tlearn: 0.9057888\ttotal: 13.9s\tremaining: 9.45s\n",
      "596:\tlearn: 0.9054032\ttotal: 14s\tremaining: 9.42s\n",
      "597:\tlearn: 0.9050154\ttotal: 14s\tremaining: 9.4s\n",
      "598:\tlearn: 0.9048565\ttotal: 14s\tremaining: 9.37s\n",
      "599:\tlearn: 0.9047246\ttotal: 14s\tremaining: 9.34s\n",
      "600:\tlearn: 0.9045230\ttotal: 14s\tremaining: 9.32s\n",
      "601:\tlearn: 0.9043029\ttotal: 14.1s\tremaining: 9.29s\n",
      "602:\tlearn: 0.9040002\ttotal: 14.1s\tremaining: 9.27s\n",
      "603:\tlearn: 0.9038515\ttotal: 14.1s\tremaining: 9.24s\n",
      "604:\tlearn: 0.9034216\ttotal: 14.1s\tremaining: 9.22s\n",
      "605:\tlearn: 0.9032178\ttotal: 14.1s\tremaining: 9.19s\n",
      "606:\tlearn: 0.9025808\ttotal: 14.2s\tremaining: 9.16s\n",
      "607:\tlearn: 0.9022980\ttotal: 14.2s\tremaining: 9.14s\n",
      "608:\tlearn: 0.9019830\ttotal: 14.2s\tremaining: 9.11s\n",
      "609:\tlearn: 0.9017598\ttotal: 14.2s\tremaining: 9.09s\n",
      "610:\tlearn: 0.9014658\ttotal: 14.2s\tremaining: 9.07s\n",
      "611:\tlearn: 0.9012494\ttotal: 14.3s\tremaining: 9.04s\n",
      "612:\tlearn: 0.9011352\ttotal: 14.3s\tremaining: 9.02s\n",
      "613:\tlearn: 0.9009632\ttotal: 14.3s\tremaining: 8.99s\n",
      "614:\tlearn: 0.9007245\ttotal: 14.3s\tremaining: 8.96s\n",
      "615:\tlearn: 0.9005230\ttotal: 14.3s\tremaining: 8.94s\n",
      "616:\tlearn: 0.9002863\ttotal: 14.4s\tremaining: 8.91s\n",
      "617:\tlearn: 0.9001758\ttotal: 14.4s\tremaining: 8.89s\n",
      "618:\tlearn: 0.8999341\ttotal: 14.4s\tremaining: 8.86s\n",
      "619:\tlearn: 0.8997598\ttotal: 14.4s\tremaining: 8.84s\n",
      "620:\tlearn: 0.8994602\ttotal: 14.4s\tremaining: 8.81s\n",
      "621:\tlearn: 0.8993066\ttotal: 14.5s\tremaining: 8.79s\n",
      "622:\tlearn: 0.8991179\ttotal: 14.5s\tremaining: 8.77s\n",
      "623:\tlearn: 0.8989473\ttotal: 14.5s\tremaining: 8.74s\n",
      "624:\tlearn: 0.8988143\ttotal: 14.5s\tremaining: 8.72s\n",
      "625:\tlearn: 0.8986394\ttotal: 14.6s\tremaining: 8.69s\n",
      "626:\tlearn: 0.8985420\ttotal: 14.6s\tremaining: 8.67s\n",
      "627:\tlearn: 0.8982983\ttotal: 14.6s\tremaining: 8.64s\n",
      "628:\tlearn: 0.8981283\ttotal: 14.6s\tremaining: 8.62s\n",
      "629:\tlearn: 0.8979969\ttotal: 14.6s\tremaining: 8.59s\n",
      "630:\tlearn: 0.8977847\ttotal: 14.6s\tremaining: 8.56s\n",
      "631:\tlearn: 0.8975000\ttotal: 14.7s\tremaining: 8.54s\n",
      "632:\tlearn: 0.8973417\ttotal: 14.7s\tremaining: 8.51s\n",
      "633:\tlearn: 0.8971771\ttotal: 14.7s\tremaining: 8.49s\n",
      "634:\tlearn: 0.8969488\ttotal: 14.7s\tremaining: 8.46s\n",
      "635:\tlearn: 0.8967724\ttotal: 14.7s\tremaining: 8.44s\n",
      "636:\tlearn: 0.8966004\ttotal: 14.8s\tremaining: 8.41s\n",
      "637:\tlearn: 0.8962209\ttotal: 14.8s\tremaining: 8.39s\n",
      "638:\tlearn: 0.8959900\ttotal: 14.8s\tremaining: 8.37s\n",
      "639:\tlearn: 0.8957673\ttotal: 14.8s\tremaining: 8.35s\n",
      "640:\tlearn: 0.8955148\ttotal: 14.9s\tremaining: 8.33s\n",
      "641:\tlearn: 0.8953000\ttotal: 14.9s\tremaining: 8.3s\n",
      "642:\tlearn: 0.8949809\ttotal: 14.9s\tremaining: 8.28s\n",
      "643:\tlearn: 0.8948534\ttotal: 14.9s\tremaining: 8.26s\n",
      "644:\tlearn: 0.8945115\ttotal: 15s\tremaining: 8.24s\n",
      "645:\tlearn: 0.8942558\ttotal: 15s\tremaining: 8.22s\n",
      "646:\tlearn: 0.8940520\ttotal: 15s\tremaining: 8.2s\n",
      "647:\tlearn: 0.8938503\ttotal: 15s\tremaining: 8.17s\n",
      "648:\tlearn: 0.8936651\ttotal: 15.1s\tremaining: 8.15s\n",
      "649:\tlearn: 0.8934517\ttotal: 15.1s\tremaining: 8.13s\n",
      "650:\tlearn: 0.8932861\ttotal: 15.1s\tremaining: 8.11s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "651:\tlearn: 0.8931407\ttotal: 15.2s\tremaining: 8.09s\n",
      "652:\tlearn: 0.8929504\ttotal: 15.2s\tremaining: 8.07s\n",
      "653:\tlearn: 0.8927449\ttotal: 15.2s\tremaining: 8.04s\n",
      "654:\tlearn: 0.8926307\ttotal: 15.2s\tremaining: 8.02s\n",
      "655:\tlearn: 0.8922137\ttotal: 15.3s\tremaining: 8s\n",
      "656:\tlearn: 0.8918407\ttotal: 15.3s\tremaining: 7.98s\n",
      "657:\tlearn: 0.8914684\ttotal: 15.3s\tremaining: 7.96s\n",
      "658:\tlearn: 0.8911421\ttotal: 15.3s\tremaining: 7.93s\n",
      "659:\tlearn: 0.8909914\ttotal: 15.4s\tremaining: 7.91s\n",
      "660:\tlearn: 0.8908645\ttotal: 15.4s\tremaining: 7.89s\n",
      "661:\tlearn: 0.8907608\ttotal: 15.4s\tremaining: 7.87s\n",
      "662:\tlearn: 0.8906599\ttotal: 15.4s\tremaining: 7.85s\n",
      "663:\tlearn: 0.8904601\ttotal: 15.5s\tremaining: 7.82s\n",
      "664:\tlearn: 0.8903037\ttotal: 15.5s\tremaining: 7.8s\n",
      "665:\tlearn: 0.8898662\ttotal: 15.5s\tremaining: 7.78s\n",
      "666:\tlearn: 0.8896227\ttotal: 15.5s\tremaining: 7.76s\n",
      "667:\tlearn: 0.8894832\ttotal: 15.6s\tremaining: 7.73s\n",
      "668:\tlearn: 0.8892017\ttotal: 15.6s\tremaining: 7.71s\n",
      "669:\tlearn: 0.8888025\ttotal: 15.6s\tremaining: 7.69s\n",
      "670:\tlearn: 0.8886980\ttotal: 15.6s\tremaining: 7.67s\n",
      "671:\tlearn: 0.8882963\ttotal: 15.7s\tremaining: 7.64s\n",
      "672:\tlearn: 0.8881304\ttotal: 15.7s\tremaining: 7.62s\n",
      "673:\tlearn: 0.8879775\ttotal: 15.7s\tremaining: 7.6s\n",
      "674:\tlearn: 0.8876720\ttotal: 15.7s\tremaining: 7.57s\n",
      "675:\tlearn: 0.8875074\ttotal: 15.8s\tremaining: 7.55s\n",
      "676:\tlearn: 0.8872171\ttotal: 15.8s\tremaining: 7.53s\n",
      "677:\tlearn: 0.8869377\ttotal: 15.8s\tremaining: 7.52s\n",
      "678:\tlearn: 0.8867551\ttotal: 15.9s\tremaining: 7.51s\n",
      "679:\tlearn: 0.8866103\ttotal: 15.9s\tremaining: 7.49s\n",
      "680:\tlearn: 0.8864376\ttotal: 15.9s\tremaining: 7.47s\n",
      "681:\tlearn: 0.8863281\ttotal: 16s\tremaining: 7.44s\n",
      "682:\tlearn: 0.8860507\ttotal: 16s\tremaining: 7.42s\n",
      "683:\tlearn: 0.8857210\ttotal: 16s\tremaining: 7.4s\n",
      "684:\tlearn: 0.8855227\ttotal: 16s\tremaining: 7.37s\n",
      "685:\tlearn: 0.8853700\ttotal: 16.1s\tremaining: 7.35s\n",
      "686:\tlearn: 0.8851845\ttotal: 16.1s\tremaining: 7.33s\n",
      "687:\tlearn: 0.8848560\ttotal: 16.1s\tremaining: 7.3s\n",
      "688:\tlearn: 0.8846811\ttotal: 16.1s\tremaining: 7.28s\n",
      "689:\tlearn: 0.8843250\ttotal: 16.1s\tremaining: 7.25s\n",
      "690:\tlearn: 0.8840245\ttotal: 16.2s\tremaining: 7.23s\n",
      "691:\tlearn: 0.8838960\ttotal: 16.2s\tremaining: 7.2s\n",
      "692:\tlearn: 0.8837290\ttotal: 16.2s\tremaining: 7.18s\n",
      "693:\tlearn: 0.8835021\ttotal: 16.2s\tremaining: 7.16s\n",
      "694:\tlearn: 0.8831070\ttotal: 16.3s\tremaining: 7.13s\n",
      "695:\tlearn: 0.8828841\ttotal: 16.3s\tremaining: 7.11s\n",
      "696:\tlearn: 0.8827418\ttotal: 16.3s\tremaining: 7.08s\n",
      "697:\tlearn: 0.8825674\ttotal: 16.3s\tremaining: 7.06s\n",
      "698:\tlearn: 0.8822695\ttotal: 16.3s\tremaining: 7.04s\n",
      "699:\tlearn: 0.8820699\ttotal: 16.4s\tremaining: 7.01s\n",
      "700:\tlearn: 0.8818914\ttotal: 16.4s\tremaining: 6.99s\n",
      "701:\tlearn: 0.8816178\ttotal: 16.4s\tremaining: 6.96s\n",
      "702:\tlearn: 0.8814378\ttotal: 16.4s\tremaining: 6.94s\n",
      "703:\tlearn: 0.8811360\ttotal: 16.4s\tremaining: 6.91s\n",
      "704:\tlearn: 0.8807689\ttotal: 16.5s\tremaining: 6.89s\n",
      "705:\tlearn: 0.8805427\ttotal: 16.5s\tremaining: 6.87s\n",
      "706:\tlearn: 0.8803114\ttotal: 16.5s\tremaining: 6.84s\n",
      "707:\tlearn: 0.8800387\ttotal: 16.5s\tremaining: 6.82s\n",
      "708:\tlearn: 0.8798186\ttotal: 16.6s\tremaining: 6.8s\n",
      "709:\tlearn: 0.8797068\ttotal: 16.6s\tremaining: 6.77s\n",
      "710:\tlearn: 0.8794708\ttotal: 16.6s\tremaining: 6.75s\n",
      "711:\tlearn: 0.8791143\ttotal: 16.6s\tremaining: 6.73s\n",
      "712:\tlearn: 0.8788305\ttotal: 16.7s\tremaining: 6.71s\n",
      "713:\tlearn: 0.8786069\ttotal: 16.7s\tremaining: 6.69s\n",
      "714:\tlearn: 0.8784626\ttotal: 16.7s\tremaining: 6.67s\n",
      "715:\tlearn: 0.8783818\ttotal: 16.8s\tremaining: 6.64s\n",
      "716:\tlearn: 0.8781534\ttotal: 16.8s\tremaining: 6.62s\n",
      "717:\tlearn: 0.8779422\ttotal: 16.8s\tremaining: 6.6s\n",
      "718:\tlearn: 0.8777365\ttotal: 16.8s\tremaining: 6.58s\n",
      "719:\tlearn: 0.8776038\ttotal: 16.9s\tremaining: 6.55s\n",
      "720:\tlearn: 0.8773649\ttotal: 16.9s\tremaining: 6.53s\n",
      "721:\tlearn: 0.8771027\ttotal: 16.9s\tremaining: 6.51s\n",
      "722:\tlearn: 0.8768837\ttotal: 16.9s\tremaining: 6.49s\n",
      "723:\tlearn: 0.8767461\ttotal: 17s\tremaining: 6.46s\n",
      "724:\tlearn: 0.8765246\ttotal: 17s\tremaining: 6.44s\n",
      "725:\tlearn: 0.8763650\ttotal: 17s\tremaining: 6.42s\n",
      "726:\tlearn: 0.8762042\ttotal: 17s\tremaining: 6.39s\n",
      "727:\tlearn: 0.8760607\ttotal: 17.1s\tremaining: 6.37s\n",
      "728:\tlearn: 0.8758395\ttotal: 17.1s\tremaining: 6.35s\n",
      "729:\tlearn: 0.8756722\ttotal: 17.1s\tremaining: 6.32s\n",
      "730:\tlearn: 0.8753855\ttotal: 17.1s\tremaining: 6.3s\n",
      "731:\tlearn: 0.8751729\ttotal: 17.1s\tremaining: 6.28s\n",
      "732:\tlearn: 0.8749720\ttotal: 17.2s\tremaining: 6.25s\n",
      "733:\tlearn: 0.8748020\ttotal: 17.2s\tremaining: 6.23s\n",
      "734:\tlearn: 0.8747004\ttotal: 17.2s\tremaining: 6.21s\n",
      "735:\tlearn: 0.8745381\ttotal: 17.2s\tremaining: 6.18s\n",
      "736:\tlearn: 0.8742345\ttotal: 17.3s\tremaining: 6.16s\n",
      "737:\tlearn: 0.8740283\ttotal: 17.3s\tremaining: 6.14s\n",
      "738:\tlearn: 0.8738623\ttotal: 17.3s\tremaining: 6.11s\n",
      "739:\tlearn: 0.8737141\ttotal: 17.3s\tremaining: 6.09s\n",
      "740:\tlearn: 0.8734430\ttotal: 17.4s\tremaining: 6.07s\n",
      "741:\tlearn: 0.8731112\ttotal: 17.4s\tremaining: 6.04s\n",
      "742:\tlearn: 0.8729335\ttotal: 17.4s\tremaining: 6.02s\n",
      "743:\tlearn: 0.8725919\ttotal: 17.4s\tremaining: 6s\n",
      "744:\tlearn: 0.8724066\ttotal: 17.4s\tremaining: 5.97s\n",
      "745:\tlearn: 0.8722121\ttotal: 17.5s\tremaining: 5.95s\n",
      "746:\tlearn: 0.8720301\ttotal: 17.5s\tremaining: 5.92s\n",
      "747:\tlearn: 0.8719164\ttotal: 17.5s\tremaining: 5.9s\n",
      "748:\tlearn: 0.8717403\ttotal: 17.5s\tremaining: 5.88s\n",
      "749:\tlearn: 0.8714995\ttotal: 17.6s\tremaining: 5.85s\n",
      "750:\tlearn: 0.8713255\ttotal: 17.6s\tremaining: 5.83s\n",
      "751:\tlearn: 0.8710973\ttotal: 17.6s\tremaining: 5.8s\n",
      "752:\tlearn: 0.8708616\ttotal: 17.6s\tremaining: 5.78s\n",
      "753:\tlearn: 0.8706502\ttotal: 17.6s\tremaining: 5.75s\n",
      "754:\tlearn: 0.8704906\ttotal: 17.7s\tremaining: 5.73s\n",
      "755:\tlearn: 0.8703396\ttotal: 17.7s\tremaining: 5.71s\n",
      "756:\tlearn: 0.8699973\ttotal: 17.7s\tremaining: 5.68s\n",
      "757:\tlearn: 0.8698830\ttotal: 17.7s\tremaining: 5.66s\n",
      "758:\tlearn: 0.8697246\ttotal: 17.7s\tremaining: 5.63s\n",
      "759:\tlearn: 0.8695271\ttotal: 17.8s\tremaining: 5.61s\n",
      "760:\tlearn: 0.8693739\ttotal: 17.8s\tremaining: 5.58s\n",
      "761:\tlearn: 0.8691340\ttotal: 17.8s\tremaining: 5.56s\n",
      "762:\tlearn: 0.8689531\ttotal: 17.8s\tremaining: 5.53s\n",
      "763:\tlearn: 0.8686748\ttotal: 17.8s\tremaining: 5.51s\n",
      "764:\tlearn: 0.8683277\ttotal: 17.9s\tremaining: 5.49s\n",
      "765:\tlearn: 0.8681497\ttotal: 17.9s\tremaining: 5.46s\n",
      "766:\tlearn: 0.8680056\ttotal: 17.9s\tremaining: 5.44s\n",
      "767:\tlearn: 0.8677874\ttotal: 17.9s\tremaining: 5.42s\n",
      "768:\tlearn: 0.8676139\ttotal: 17.9s\tremaining: 5.39s\n",
      "769:\tlearn: 0.8674177\ttotal: 18s\tremaining: 5.37s\n",
      "770:\tlearn: 0.8672485\ttotal: 18s\tremaining: 5.34s\n",
      "771:\tlearn: 0.8670956\ttotal: 18s\tremaining: 5.32s\n",
      "772:\tlearn: 0.8667783\ttotal: 18s\tremaining: 5.3s\n",
      "773:\tlearn: 0.8666439\ttotal: 18.1s\tremaining: 5.27s\n",
      "774:\tlearn: 0.8665216\ttotal: 18.1s\tremaining: 5.25s\n",
      "775:\tlearn: 0.8663134\ttotal: 18.1s\tremaining: 5.22s\n",
      "776:\tlearn: 0.8661900\ttotal: 18.1s\tremaining: 5.2s\n",
      "777:\tlearn: 0.8660535\ttotal: 18.1s\tremaining: 5.18s\n",
      "778:\tlearn: 0.8658452\ttotal: 18.2s\tremaining: 5.15s\n",
      "779:\tlearn: 0.8655855\ttotal: 18.2s\tremaining: 5.13s\n",
      "780:\tlearn: 0.8653672\ttotal: 18.2s\tremaining: 5.11s\n",
      "781:\tlearn: 0.8651921\ttotal: 18.3s\tremaining: 5.09s\n",
      "782:\tlearn: 0.8647833\ttotal: 18.3s\tremaining: 5.07s\n",
      "783:\tlearn: 0.8645855\ttotal: 18.3s\tremaining: 5.04s\n",
      "784:\tlearn: 0.8644826\ttotal: 18.3s\tremaining: 5.02s\n",
      "785:\tlearn: 0.8643126\ttotal: 18.3s\tremaining: 4.99s\n",
      "786:\tlearn: 0.8641955\ttotal: 18.4s\tremaining: 4.97s\n",
      "787:\tlearn: 0.8640402\ttotal: 18.4s\tremaining: 4.95s\n",
      "788:\tlearn: 0.8638919\ttotal: 18.4s\tremaining: 4.92s\n",
      "789:\tlearn: 0.8637928\ttotal: 18.4s\tremaining: 4.9s\n",
      "790:\tlearn: 0.8634684\ttotal: 18.5s\tremaining: 4.88s\n",
      "791:\tlearn: 0.8633380\ttotal: 18.5s\tremaining: 4.85s\n",
      "792:\tlearn: 0.8631799\ttotal: 18.5s\tremaining: 4.83s\n",
      "793:\tlearn: 0.8630062\ttotal: 18.5s\tremaining: 4.81s\n",
      "794:\tlearn: 0.8628039\ttotal: 18.6s\tremaining: 4.78s\n",
      "795:\tlearn: 0.8625961\ttotal: 18.6s\tremaining: 4.76s\n",
      "796:\tlearn: 0.8624387\ttotal: 18.6s\tremaining: 4.74s\n",
      "797:\tlearn: 0.8621238\ttotal: 18.6s\tremaining: 4.71s\n",
      "798:\tlearn: 0.8620249\ttotal: 18.7s\tremaining: 4.69s\n",
      "799:\tlearn: 0.8618811\ttotal: 18.7s\tremaining: 4.67s\n",
      "800:\tlearn: 0.8616097\ttotal: 18.7s\tremaining: 4.65s\n",
      "801:\tlearn: 0.8614241\ttotal: 18.7s\tremaining: 4.62s\n",
      "802:\tlearn: 0.8612596\ttotal: 18.7s\tremaining: 4.6s\n",
      "803:\tlearn: 0.8611190\ttotal: 18.8s\tremaining: 4.58s\n",
      "804:\tlearn: 0.8610273\ttotal: 18.8s\tremaining: 4.55s\n",
      "805:\tlearn: 0.8607828\ttotal: 18.8s\tremaining: 4.53s\n",
      "806:\tlearn: 0.8603833\ttotal: 18.8s\tremaining: 4.5s\n",
      "807:\tlearn: 0.8601904\ttotal: 18.9s\tremaining: 4.48s\n",
      "808:\tlearn: 0.8600085\ttotal: 18.9s\tremaining: 4.46s\n",
      "809:\tlearn: 0.8597863\ttotal: 18.9s\tremaining: 4.44s\n",
      "810:\tlearn: 0.8595945\ttotal: 18.9s\tremaining: 4.41s\n",
      "811:\tlearn: 0.8594253\ttotal: 19s\tremaining: 4.39s\n",
      "812:\tlearn: 0.8593176\ttotal: 19s\tremaining: 4.37s\n",
      "813:\tlearn: 0.8590060\ttotal: 19s\tremaining: 4.34s\n",
      "814:\tlearn: 0.8587130\ttotal: 19s\tremaining: 4.32s\n",
      "815:\tlearn: 0.8584854\ttotal: 19s\tremaining: 4.29s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "816:\tlearn: 0.8582319\ttotal: 19.1s\tremaining: 4.27s\n",
      "817:\tlearn: 0.8580545\ttotal: 19.1s\tremaining: 4.25s\n",
      "818:\tlearn: 0.8579261\ttotal: 19.1s\tremaining: 4.22s\n",
      "819:\tlearn: 0.8576346\ttotal: 19.1s\tremaining: 4.2s\n",
      "820:\tlearn: 0.8575122\ttotal: 19.2s\tremaining: 4.18s\n",
      "821:\tlearn: 0.8573658\ttotal: 19.2s\tremaining: 4.15s\n",
      "822:\tlearn: 0.8572540\ttotal: 19.2s\tremaining: 4.13s\n",
      "823:\tlearn: 0.8569409\ttotal: 19.2s\tremaining: 4.11s\n",
      "824:\tlearn: 0.8566792\ttotal: 19.2s\tremaining: 4.08s\n",
      "825:\tlearn: 0.8564978\ttotal: 19.3s\tremaining: 4.06s\n",
      "826:\tlearn: 0.8562763\ttotal: 19.3s\tremaining: 4.04s\n",
      "827:\tlearn: 0.8560856\ttotal: 19.3s\tremaining: 4.01s\n",
      "828:\tlearn: 0.8558119\ttotal: 19.3s\tremaining: 3.99s\n",
      "829:\tlearn: 0.8556240\ttotal: 19.4s\tremaining: 3.96s\n",
      "830:\tlearn: 0.8555293\ttotal: 19.4s\tremaining: 3.94s\n",
      "831:\tlearn: 0.8554134\ttotal: 19.4s\tremaining: 3.92s\n",
      "832:\tlearn: 0.8552534\ttotal: 19.4s\tremaining: 3.89s\n",
      "833:\tlearn: 0.8551265\ttotal: 19.4s\tremaining: 3.87s\n",
      "834:\tlearn: 0.8549774\ttotal: 19.5s\tremaining: 3.85s\n",
      "835:\tlearn: 0.8548278\ttotal: 19.5s\tremaining: 3.83s\n",
      "836:\tlearn: 0.8547473\ttotal: 19.5s\tremaining: 3.8s\n",
      "837:\tlearn: 0.8545304\ttotal: 19.6s\tremaining: 3.78s\n",
      "838:\tlearn: 0.8542246\ttotal: 19.6s\tremaining: 3.76s\n",
      "839:\tlearn: 0.8540571\ttotal: 19.6s\tremaining: 3.73s\n",
      "840:\tlearn: 0.8539297\ttotal: 19.6s\tremaining: 3.71s\n",
      "841:\tlearn: 0.8537872\ttotal: 19.6s\tremaining: 3.68s\n",
      "842:\tlearn: 0.8536707\ttotal: 19.7s\tremaining: 3.66s\n",
      "843:\tlearn: 0.8534577\ttotal: 19.7s\tremaining: 3.64s\n",
      "844:\tlearn: 0.8533271\ttotal: 19.7s\tremaining: 3.61s\n",
      "845:\tlearn: 0.8531908\ttotal: 19.7s\tremaining: 3.59s\n",
      "846:\tlearn: 0.8530525\ttotal: 19.7s\tremaining: 3.57s\n",
      "847:\tlearn: 0.8529255\ttotal: 19.8s\tremaining: 3.54s\n",
      "848:\tlearn: 0.8528049\ttotal: 19.8s\tremaining: 3.52s\n",
      "849:\tlearn: 0.8526021\ttotal: 19.8s\tremaining: 3.5s\n",
      "850:\tlearn: 0.8524060\ttotal: 19.8s\tremaining: 3.47s\n",
      "851:\tlearn: 0.8522292\ttotal: 19.8s\tremaining: 3.45s\n",
      "852:\tlearn: 0.8520413\ttotal: 19.9s\tremaining: 3.42s\n",
      "853:\tlearn: 0.8518623\ttotal: 19.9s\tremaining: 3.4s\n",
      "854:\tlearn: 0.8516945\ttotal: 19.9s\tremaining: 3.38s\n",
      "855:\tlearn: 0.8514956\ttotal: 19.9s\tremaining: 3.35s\n",
      "856:\tlearn: 0.8514113\ttotal: 20s\tremaining: 3.33s\n",
      "857:\tlearn: 0.8509812\ttotal: 20s\tremaining: 3.31s\n",
      "858:\tlearn: 0.8508915\ttotal: 20s\tremaining: 3.28s\n",
      "859:\tlearn: 0.8506610\ttotal: 20s\tremaining: 3.26s\n",
      "860:\tlearn: 0.8504145\ttotal: 20s\tremaining: 3.23s\n",
      "861:\tlearn: 0.8502373\ttotal: 20.1s\tremaining: 3.21s\n",
      "862:\tlearn: 0.8501008\ttotal: 20.1s\tremaining: 3.19s\n",
      "863:\tlearn: 0.8499147\ttotal: 20.1s\tremaining: 3.16s\n",
      "864:\tlearn: 0.8496843\ttotal: 20.1s\tremaining: 3.14s\n",
      "865:\tlearn: 0.8494260\ttotal: 20.1s\tremaining: 3.12s\n",
      "866:\tlearn: 0.8492655\ttotal: 20.2s\tremaining: 3.09s\n",
      "867:\tlearn: 0.8491050\ttotal: 20.2s\tremaining: 3.07s\n",
      "868:\tlearn: 0.8490094\ttotal: 20.2s\tremaining: 3.05s\n",
      "869:\tlearn: 0.8487757\ttotal: 20.2s\tremaining: 3.02s\n",
      "870:\tlearn: 0.8486940\ttotal: 20.3s\tremaining: 3s\n",
      "871:\tlearn: 0.8485015\ttotal: 20.3s\tremaining: 2.98s\n",
      "872:\tlearn: 0.8483189\ttotal: 20.3s\tremaining: 2.96s\n",
      "873:\tlearn: 0.8482300\ttotal: 20.3s\tremaining: 2.93s\n",
      "874:\tlearn: 0.8480625\ttotal: 20.4s\tremaining: 2.91s\n",
      "875:\tlearn: 0.8478114\ttotal: 20.4s\tremaining: 2.88s\n",
      "876:\tlearn: 0.8475275\ttotal: 20.4s\tremaining: 2.86s\n",
      "877:\tlearn: 0.8473849\ttotal: 20.4s\tremaining: 2.84s\n",
      "878:\tlearn: 0.8472496\ttotal: 20.5s\tremaining: 2.82s\n",
      "879:\tlearn: 0.8470096\ttotal: 20.5s\tremaining: 2.79s\n",
      "880:\tlearn: 0.8467907\ttotal: 20.5s\tremaining: 2.77s\n",
      "881:\tlearn: 0.8464524\ttotal: 20.5s\tremaining: 2.75s\n",
      "882:\tlearn: 0.8462349\ttotal: 20.6s\tremaining: 2.72s\n",
      "883:\tlearn: 0.8460605\ttotal: 20.6s\tremaining: 2.7s\n",
      "884:\tlearn: 0.8459070\ttotal: 20.6s\tremaining: 2.68s\n",
      "885:\tlearn: 0.8456559\ttotal: 20.6s\tremaining: 2.65s\n",
      "886:\tlearn: 0.8453380\ttotal: 20.7s\tremaining: 2.63s\n",
      "887:\tlearn: 0.8448708\ttotal: 20.7s\tremaining: 2.61s\n",
      "888:\tlearn: 0.8447140\ttotal: 20.7s\tremaining: 2.58s\n",
      "889:\tlearn: 0.8445174\ttotal: 20.7s\tremaining: 2.56s\n",
      "890:\tlearn: 0.8442661\ttotal: 20.8s\tremaining: 2.54s\n",
      "891:\tlearn: 0.8438736\ttotal: 20.8s\tremaining: 2.52s\n",
      "892:\tlearn: 0.8436586\ttotal: 20.8s\tremaining: 2.49s\n",
      "893:\tlearn: 0.8434213\ttotal: 20.8s\tremaining: 2.47s\n",
      "894:\tlearn: 0.8431284\ttotal: 20.8s\tremaining: 2.44s\n",
      "895:\tlearn: 0.8429407\ttotal: 20.9s\tremaining: 2.42s\n",
      "896:\tlearn: 0.8428398\ttotal: 20.9s\tremaining: 2.4s\n",
      "897:\tlearn: 0.8426925\ttotal: 20.9s\tremaining: 2.37s\n",
      "898:\tlearn: 0.8425800\ttotal: 20.9s\tremaining: 2.35s\n",
      "899:\tlearn: 0.8424052\ttotal: 20.9s\tremaining: 2.33s\n",
      "900:\tlearn: 0.8421342\ttotal: 21s\tremaining: 2.3s\n",
      "901:\tlearn: 0.8418426\ttotal: 21s\tremaining: 2.28s\n",
      "902:\tlearn: 0.8417240\ttotal: 21s\tremaining: 2.26s\n",
      "903:\tlearn: 0.8414460\ttotal: 21s\tremaining: 2.23s\n",
      "904:\tlearn: 0.8412291\ttotal: 21.1s\tremaining: 2.21s\n",
      "905:\tlearn: 0.8410587\ttotal: 21.1s\tremaining: 2.19s\n",
      "906:\tlearn: 0.8408151\ttotal: 21.1s\tremaining: 2.16s\n",
      "907:\tlearn: 0.8404463\ttotal: 21.1s\tremaining: 2.14s\n",
      "908:\tlearn: 0.8403055\ttotal: 21.1s\tremaining: 2.12s\n",
      "909:\tlearn: 0.8401413\ttotal: 21.2s\tremaining: 2.09s\n",
      "910:\tlearn: 0.8399469\ttotal: 21.2s\tremaining: 2.07s\n",
      "911:\tlearn: 0.8396937\ttotal: 21.2s\tremaining: 2.05s\n",
      "912:\tlearn: 0.8394499\ttotal: 21.2s\tremaining: 2.02s\n",
      "913:\tlearn: 0.8393196\ttotal: 21.2s\tremaining: 2s\n",
      "914:\tlearn: 0.8391988\ttotal: 21.3s\tremaining: 1.98s\n",
      "915:\tlearn: 0.8389752\ttotal: 21.3s\tremaining: 1.95s\n",
      "916:\tlearn: 0.8387663\ttotal: 21.3s\tremaining: 1.93s\n",
      "917:\tlearn: 0.8386342\ttotal: 21.3s\tremaining: 1.91s\n",
      "918:\tlearn: 0.8384715\ttotal: 21.3s\tremaining: 1.88s\n",
      "919:\tlearn: 0.8382847\ttotal: 21.4s\tremaining: 1.86s\n",
      "920:\tlearn: 0.8380269\ttotal: 21.4s\tremaining: 1.83s\n",
      "921:\tlearn: 0.8378501\ttotal: 21.4s\tremaining: 1.81s\n",
      "922:\tlearn: 0.8376200\ttotal: 21.4s\tremaining: 1.79s\n",
      "923:\tlearn: 0.8373816\ttotal: 21.5s\tremaining: 1.76s\n",
      "924:\tlearn: 0.8372553\ttotal: 21.5s\tremaining: 1.74s\n",
      "925:\tlearn: 0.8370739\ttotal: 21.5s\tremaining: 1.72s\n",
      "926:\tlearn: 0.8367467\ttotal: 21.5s\tremaining: 1.69s\n",
      "927:\tlearn: 0.8366268\ttotal: 21.5s\tremaining: 1.67s\n",
      "928:\tlearn: 0.8364745\ttotal: 21.6s\tremaining: 1.65s\n",
      "929:\tlearn: 0.8363132\ttotal: 21.6s\tremaining: 1.63s\n",
      "930:\tlearn: 0.8360317\ttotal: 21.6s\tremaining: 1.6s\n",
      "931:\tlearn: 0.8358651\ttotal: 21.6s\tremaining: 1.58s\n",
      "932:\tlearn: 0.8354264\ttotal: 21.7s\tremaining: 1.55s\n",
      "933:\tlearn: 0.8353501\ttotal: 21.7s\tremaining: 1.53s\n",
      "934:\tlearn: 0.8352317\ttotal: 21.7s\tremaining: 1.51s\n",
      "935:\tlearn: 0.8351000\ttotal: 21.7s\tremaining: 1.49s\n",
      "936:\tlearn: 0.8349482\ttotal: 21.7s\tremaining: 1.46s\n",
      "937:\tlearn: 0.8348146\ttotal: 21.8s\tremaining: 1.44s\n",
      "938:\tlearn: 0.8346520\ttotal: 21.8s\tremaining: 1.42s\n",
      "939:\tlearn: 0.8345070\ttotal: 21.8s\tremaining: 1.39s\n",
      "940:\tlearn: 0.8343811\ttotal: 21.8s\tremaining: 1.37s\n",
      "941:\tlearn: 0.8341437\ttotal: 21.9s\tremaining: 1.35s\n",
      "942:\tlearn: 0.8339916\ttotal: 21.9s\tremaining: 1.32s\n",
      "943:\tlearn: 0.8338136\ttotal: 21.9s\tremaining: 1.3s\n",
      "944:\tlearn: 0.8336663\ttotal: 21.9s\tremaining: 1.28s\n",
      "945:\tlearn: 0.8334931\ttotal: 22s\tremaining: 1.25s\n",
      "946:\tlearn: 0.8333568\ttotal: 22s\tremaining: 1.23s\n",
      "947:\tlearn: 0.8332559\ttotal: 22s\tremaining: 1.21s\n",
      "948:\tlearn: 0.8330817\ttotal: 22s\tremaining: 1.18s\n",
      "949:\tlearn: 0.8328705\ttotal: 22s\tremaining: 1.16s\n",
      "950:\tlearn: 0.8327721\ttotal: 22.1s\tremaining: 1.14s\n",
      "951:\tlearn: 0.8326249\ttotal: 22.1s\tremaining: 1.11s\n",
      "952:\tlearn: 0.8325303\ttotal: 22.1s\tremaining: 1.09s\n",
      "953:\tlearn: 0.8324251\ttotal: 22.1s\tremaining: 1.07s\n",
      "954:\tlearn: 0.8323251\ttotal: 22.2s\tremaining: 1.04s\n",
      "955:\tlearn: 0.8321626\ttotal: 22.2s\tremaining: 1.02s\n",
      "956:\tlearn: 0.8320371\ttotal: 22.2s\tremaining: 998ms\n",
      "957:\tlearn: 0.8318314\ttotal: 22.2s\tremaining: 974ms\n",
      "958:\tlearn: 0.8315843\ttotal: 22.3s\tremaining: 951ms\n",
      "959:\tlearn: 0.8314148\ttotal: 22.3s\tremaining: 928ms\n",
      "960:\tlearn: 0.8313110\ttotal: 22.3s\tremaining: 905ms\n",
      "961:\tlearn: 0.8312090\ttotal: 22.3s\tremaining: 881ms\n",
      "962:\tlearn: 0.8309265\ttotal: 22.3s\tremaining: 858ms\n",
      "963:\tlearn: 0.8307865\ttotal: 22.4s\tremaining: 835ms\n",
      "964:\tlearn: 0.8305616\ttotal: 22.4s\tremaining: 811ms\n",
      "965:\tlearn: 0.8303767\ttotal: 22.4s\tremaining: 788ms\n",
      "966:\tlearn: 0.8301804\ttotal: 22.4s\tremaining: 765ms\n",
      "967:\tlearn: 0.8300142\ttotal: 22.4s\tremaining: 742ms\n",
      "968:\tlearn: 0.8299247\ttotal: 22.5s\tremaining: 718ms\n",
      "969:\tlearn: 0.8297080\ttotal: 22.5s\tremaining: 696ms\n",
      "970:\tlearn: 0.8295495\ttotal: 22.5s\tremaining: 672ms\n",
      "971:\tlearn: 0.8293755\ttotal: 22.5s\tremaining: 649ms\n",
      "972:\tlearn: 0.8292352\ttotal: 22.6s\tremaining: 626ms\n",
      "973:\tlearn: 0.8289485\ttotal: 22.6s\tremaining: 603ms\n",
      "974:\tlearn: 0.8288377\ttotal: 22.6s\tremaining: 579ms\n",
      "975:\tlearn: 0.8287273\ttotal: 22.6s\tremaining: 556ms\n",
      "976:\tlearn: 0.8286093\ttotal: 22.6s\tremaining: 533ms\n",
      "977:\tlearn: 0.8283560\ttotal: 22.7s\tremaining: 510ms\n",
      "978:\tlearn: 0.8281101\ttotal: 22.7s\tremaining: 486ms\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "979:\tlearn: 0.8279538\ttotal: 22.7s\tremaining: 463ms\n",
      "980:\tlearn: 0.8278237\ttotal: 22.7s\tremaining: 440ms\n",
      "981:\tlearn: 0.8277115\ttotal: 22.7s\tremaining: 417ms\n",
      "982:\tlearn: 0.8275455\ttotal: 22.8s\tremaining: 394ms\n",
      "983:\tlearn: 0.8274036\ttotal: 22.8s\tremaining: 371ms\n",
      "984:\tlearn: 0.8272325\ttotal: 22.8s\tremaining: 347ms\n",
      "985:\tlearn: 0.8270935\ttotal: 22.8s\tremaining: 324ms\n",
      "986:\tlearn: 0.8266488\ttotal: 22.8s\tremaining: 301ms\n",
      "987:\tlearn: 0.8265203\ttotal: 22.9s\tremaining: 278ms\n",
      "988:\tlearn: 0.8263941\ttotal: 22.9s\tremaining: 255ms\n",
      "989:\tlearn: 0.8261462\ttotal: 22.9s\tremaining: 231ms\n",
      "990:\tlearn: 0.8260152\ttotal: 22.9s\tremaining: 208ms\n",
      "991:\tlearn: 0.8255935\ttotal: 22.9s\tremaining: 185ms\n",
      "992:\tlearn: 0.8254014\ttotal: 23s\tremaining: 162ms\n",
      "993:\tlearn: 0.8251952\ttotal: 23s\tremaining: 139ms\n",
      "994:\tlearn: 0.8250403\ttotal: 23s\tremaining: 116ms\n",
      "995:\tlearn: 0.8249514\ttotal: 23s\tremaining: 92.5ms\n",
      "996:\tlearn: 0.8248548\ttotal: 23s\tremaining: 69.3ms\n",
      "997:\tlearn: 0.8246792\ttotal: 23.1s\tremaining: 46.2ms\n",
      "998:\tlearn: 0.8245898\ttotal: 23.1s\tremaining: 23.1ms\n",
      "999:\tlearn: 0.8244262\ttotal: 23.1s\tremaining: 0us\n",
      "confusion matrix:\n",
      " [[ 36   0   0   8  77   6   0]\n",
      " [  7   0   0   1  16   0   0]\n",
      " [  0   0   2   0  12   0   0]\n",
      " [  3   0   0  99 151   1   1]\n",
      " [  4   0   2  26 532   6   0]\n",
      " [  1   0   0   2  52  81   0]\n",
      " [  0   0   0  14  35   0   3]]\n",
      "accuracy score =  0.6392190152801358\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.71      0.28      0.40       127\n",
      "    disgust        0.00      0.00      0.00        24\n",
      "       fear        0.50      0.14      0.22        14\n",
      "        joy        0.66      0.39      0.49       255\n",
      "     others        0.61      0.93      0.74       570\n",
      "    sadness        0.86      0.60      0.70       136\n",
      "   surprise        0.75      0.06      0.11        52\n",
      "\n",
      "    accuracy                           0.64      1178\n",
      "   macro avg       0.58      0.34      0.38      1178\n",
      "weighted avg       0.65      0.64      0.59      1178\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x20a00062cd0>"
      ]
     },
     "execution_count": 13,
     "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": [
    "#Cat Boost\n",
    "#training model\n",
    "from catboost import CatBoostClassifier\n",
    "cb = CatBoostClassifier()\n",
    "cb.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = cb.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "cba = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',accuracy_score(y_test,y_pred))\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(cb,x_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd001084",
   "metadata": {},
   "source": [
    "## Regresión Logística.\n",
    "    Es un algoritmo de aprendizaje supervisado que se utiliza para la clasificación. El modelo creado por este algoritmo se basa en la función logística. Una función logística, también denominada curva logística, es una curva sigmoidea, donde $x_0$ es el valor  del punto medio de $x$ de la sigmoide, $M$ es el valor máximo de la curva y $s$ es la es la inclinación de la curva.\n",
    "    $$ f(x) =  \\dfrac{M}{1 + e^{-k(x-x_0)}}$$\n",
    "    Esta función logística convierte la salida de una regresión lineal en clases. Una mayor linealidad entre la característica y la variable objetivo contribuye a un mejor rendimiento del modelo de regresión logística."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "2fd143da",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "confusion matrix:\n",
      " [[ 58   7   2  10  32  11   7]\n",
      " [  9   7   0   1   2   2   3]\n",
      " [  1   1   5   0   3   0   4]\n",
      " [ 13   8   2 131  71   8  22]\n",
      " [ 47  16   8  93 333  23  50]\n",
      " [  6   3   2   4  21  92   8]\n",
      " [  4   0   4  12  12   2  18]]\n",
      "accuracy score =  0.5466893039049237\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.42      0.46      0.44       127\n",
      "    disgust        0.17      0.29      0.21        24\n",
      "       fear        0.22      0.36      0.27        14\n",
      "        joy        0.52      0.51      0.52       255\n",
      "     others        0.70      0.58      0.64       570\n",
      "    sadness        0.67      0.68      0.67       136\n",
      "   surprise        0.16      0.35      0.22        52\n",
      "\n",
      "    accuracy                           0.55      1178\n",
      "   macro avg       0.41      0.46      0.42      1178\n",
      "weighted avg       0.59      0.55      0.56      1178\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\jabel\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\_logistic.py:763: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  n_iter_i = _check_optimize_result(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x21c60b42e80>"
      ]
     },
     "execution_count": 9,
     "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": [
    "#Logistic Regression\n",
    "#training model\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "lr = LogisticRegression(multi_class='multinomial', solver='lbfgs', penalty='l2', C=1.0, class_weight='balanced')\n",
    "lr.fit(x_train,y_train)\n",
    "\n",
    "#getting confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred = lr.predict(x_test)\n",
    "cm = confusion_matrix(y_test,y_pred)\n",
    "print('confusion matrix:\\n',cm)\n",
    "\n",
    "#checking accuracy\n",
    "from sklearn.metrics import accuracy_score\n",
    "lra = accuracy_score(y_test,y_pred)\n",
    "print('accuracy score = ',lra)\n",
    "\n",
    "#cheking metrics and plot matrix confussion\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(lr,x_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47956949",
   "metadata": {},
   "source": [
    "##  MLPercertron. \n",
    "El percertrón es la red neuronal más simple usada para la clasificación de un tipo especial de patrones, los linealmente separables (es decir, patrones que se encuentran a ambos lados de un hiperplano). Consiste de una neurona con pesos sinápticos y umbral ajustables. El algoritmo usado para ajustar los parámetros libres de esta red neuronal apareció por primera vez en un procedimiento de aprendizaje desarrollado por Rosenblatt \\cite{mt7} para su modelo de perceptrón del cerebro. Rosenblatt demostró que si los patrones usados para entrenar el perceptrón son sacados de dos clases linealmente separables, entonces el algoritmo del perceptrón converge y toma como superficie de decisión un hiperplano. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "e8f78a65",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The accuracy of the MLP is 0.5373514431239389\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "      anger        0.55      0.38      0.45       127\n",
      "    disgust        0.25      0.04      0.07        24\n",
      "       fear        0.50      0.07      0.12        14\n",
      "        joy        0.53      0.40      0.45       255\n",
      "     others        0.62      0.84      0.72       570\n",
      "    sadness        0.79      0.67      0.73       136\n",
      "   surprise        0.60      0.12      0.19        52\n",
      "\n",
      "    accuracy                           0.62      1178\n",
      "   macro avg       0.55      0.36      0.39      1178\n",
      "weighted avg       0.61      0.62      0.59      1178\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x20a7144bc10>"
      ]
     },
     "execution_count": 15,
     "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": [
    "#MLPercentron\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "mlp = MLPClassifier(random_state=1, max_iter=300)\n",
    "mlp.fit(x_train, y_train)\n",
    "score_mlp = mlp.score(x_test,y_test)\n",
    "print('The accuracy of the MLP is', score_mlp)\n",
    "print(classification_report(y_test, y_pred))\n",
    "plot_confusion_matrix(mlp,x_test,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "59113e8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Plotting the Model Accuracies')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "plt.figure(figsize= (8,7))\n",
    "\n",
    "ac = [lra,knna,sva,sva2,nba,dta,rfa,cba]\n",
    "name = ['Logistic Regression','knn','svm','Kernel Svm','Naive Bayes','Decision Tree', 'Random Forest','Cat Boost']\n",
    "sns.barplot(x = ac,y = name,palette='pastel')\n",
    "plt.title(\"Plotting the Model Accuracies\", fontsize=16, fontweight=\"bold\")"
   ]
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}