0. Trabajo Final MIA/Notebook Actividad Final Eduardo Carrasco.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "afd0f184-d796-4b88-907c-8797e7a6790b",
   "metadata": {
    "tags": []
   },
   "source": [
    "## **Eduardo Carrasco Vidal** <img src=\"img/logo.png\" align=\"right\" style=\"width: 120px;\"/>\n",
    "**Magister en Inteligencia Artificial, Universidad Adolfo Ibáñez.**\n",
    "\n",
    "**Profesor:** Adrián Soto Suárez.\n",
    "**Curso:** Aprendizaje de Máquinas (Machine Learning).\n",
    "\n",
    "Enlace al repositorio del alumno en [GitHub](https://github.com/educarrascov/MIA_MachineLearning) _@educarrascov_\n",
    "\n",
    "![Python](https://img.shields.io/badge/python-%2314354C.svg) \n",
    "\n",
    "\n",
    "# Actividad Aprendizaje Automático\n",
    "\n",
    "En este _notebook_ vamos a presentar la actividad final del curso aprendizaje automático. Se divide en 3 partes:\n",
    "\n",
    "- Evaluación de desempeño de un clasificador.\n",
    "- Clustering y reducción de dimensionalidad.\n",
    "- Clasificación con Random Forests.\n",
    "\n",
    "Cada una de estas partes vale **2 puntos**. La nota final será el **número de puntos + 1**."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b2f1e35-cf94-4c2b-a3c6-6fa5420eb7c5",
   "metadata": {},
   "source": [
    "## 1. Evaluando el clasificador Naive Bayes\\[2 pts\\]\n",
    "\n",
    "En esta pregunta debes evaluar el clasificador **Naive Bayes** presentado en clases, que predecía si un correo era Spam o no. Para evaluarlo se pide que calcules:\n",
    "\n",
    "- El _accuracy_.\n",
    "- _Precision_ y _recall_.\n",
    "- _F1-Score_.\n",
    "- Curva de _Precision vs Recall_.\n",
    "- Curva ROC.\n",
    "\n",
    "Luego, en base a tus resultados, debes explicar que tan bueno crees que es el clasificador y si crees que se puede mejorar."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a05a905b-48f6-4e67-9ca6-e7843ca8fdfc",
   "metadata": {},
   "source": [
    "Primero se crea el entorno para efectuar la clasificación binaria de los correos: spam o no spam. Para lo anterior, se utilizará el calsificador de Naive Bayes.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 365,
   "id": "30f1e319-686b-4ad8-a08b-2b02fb5f1c64",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB()"
      ]
     },
     "execution_count": 365,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.model_selection import train_test_split\n",
    "import glob\n",
    "\n",
    "# Recuerda ubicar los mails junto a este _notebook_\n",
    "PATH_MAILS = 'mails-raw/*/*'\n",
    "\n",
    "mails = []\n",
    "answers = []\n",
    "\n",
    "# entrenar el clasificador Naive bayes\n",
    "\n",
    "for filename in glob.glob(PATH_MAILS):\n",
    "    # Checkeamos si es spam o no\n",
    "    is_spam = 'ham' not in filename\n",
    "    \n",
    "    # Ignoramos los posibles errores al abrir un archivo\n",
    "    with open(filename, errors='ignore') as mail_file:\n",
    "        for line in mail_file:\n",
    "            if line.startswith(\"Subject:\"):\n",
    "                # Hacemos strip desde la izquierda\n",
    "                subject = line.lstrip(\"Subject: \")\n",
    "                mails.append(subject.strip())\n",
    "                if is_spam:\n",
    "                    answers.append(1)\n",
    "                else:\n",
    "                    answers.append(0)\n",
    "                break\n",
    "                \n",
    "# Creamos los _datasets_\n",
    "X_train, X_test, y_train, y_test = train_test_split(mails, answers)\n",
    "\n",
    "# Count vectorizer para representar las features\n",
    "vect = CountVectorizer() # contar palabras en una dimensión\n",
    "vect.fit(X_train)\n",
    "\n",
    "# Creamos nuestra matriz para el _dataset_ de entrenamiento\n",
    "X_train_vect = vect.transform(X_train)\n",
    "# Creamos nuestra matriz para el _dataset_ de rueba\n",
    "X_test_vect = vect.transform(X_test) # Notar que estamos transformando con el fit del dataset de entrenamiento\n",
    "\n",
    "# Instanciamos y entrenamos el modelo\n",
    "nb= MultinomialNB()\n",
    "nb.fit(X_train_vect, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 366,
   "id": "506a485e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "scipy.sparse.csr.csr_matrix"
      ]
     },
     "execution_count": 366,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(X_train_vect) # para ver el tipo de X_train_vect, entrega una matriz sparse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 367,
   "id": "2fdef9b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  (0, 652)\t1\n",
      "  (0, 1333)\t1\n",
      "  (0, 2511)\t1\n",
      "  (0, 3331)\t1\n",
      "  (0, 3419)\t1\n",
      "  (1, 1151)\t1\n",
      "  (1, 2710)\t1\n",
      "  (1, 2854)\t1\n",
      "  (1, 2942)\t1\n",
      "  (1, 3139)\t1\n",
      "  (1, 3443)\t1\n",
      "  (2, 360)\t1\n",
      "  (2, 384)\t1\n",
      "  (2, 414)\t1\n",
      "  (2, 539)\t1\n",
      "  (2, 1703)\t1\n",
      "  (2, 2696)\t1\n",
      "  (2, 2823)\t1\n",
      "  (2, 3418)\t1\n",
      "  (2, 3719)\t1\n",
      "  (3, 69)\t1\n",
      "  (3, 1517)\t1\n",
      "  (3, 1778)\t1\n",
      "  (3, 1826)\t1\n",
      "  (3, 2289)\t1\n",
      "  :\t:\n",
      "  (2470, 1768)\t1\n",
      "  (2470, 2410)\t1\n",
      "  (2470, 3302)\t1\n",
      "  (2470, 3540)\t1\n",
      "  (2471, 511)\t1\n",
      "  (2471, 2365)\t1\n",
      "  (2471, 2710)\t2\n",
      "  (2471, 2954)\t1\n",
      "  (2471, 3550)\t1\n",
      "  (2471, 3577)\t1\n",
      "  (2472, 333)\t1\n",
      "  (2472, 2710)\t1\n",
      "  (2472, 3700)\t1\n",
      "  (2473, 45)\t1\n",
      "  (2473, 127)\t1\n",
      "  (2473, 767)\t1\n",
      "  (2473, 1263)\t1\n",
      "  (2473, 1505)\t1\n",
      "  (2473, 1816)\t1\n",
      "  (2473, 3031)\t1\n",
      "  (2473, 3278)\t1\n",
      "  (2473, 3563)\t1\n",
      "  (2474, 2196)\t1\n",
      "  (2474, 2318)\t1\n",
      "  (2474, 2710)\t1\n"
     ]
    }
   ],
   "source": [
    "print (X_train_vect) #obtenemos un vector que informa, posición 0, palabra 438-2397 y 2741."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 368,
   "id": "e5c158eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0,\n",
       "       1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1,\n",
       "       1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0,\n",
       "       0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0,\n",
       "       0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
       "       0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1,\n",
       "       0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,\n",
       "       1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       1, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0,\n",
       "       0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0,\n",
       "       0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])"
      ]
     },
     "execution_count": 368,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nb.predict(X_test_vect) # para obtener las predicciones de los datos de prueba."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 369,
   "id": "b2200a31",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_predicted = nb.predict(X_test_vect) # con esta generamos la función de predicción que nos permitirá evaluar las métricas."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "714eb9cb-4d66-44c4-9259-4ed99d748609",
   "metadata": {},
   "source": [
    "### 1.1. Accuracy: \n",
    "Antes de efectuar el cálculo de las métricas, se efectuará un cálculo de la **matriz de confusión** con el objeto de comprender de manera gráfica los siguientes parámetros: **T**rue **P**ositive (Verdadero Positivo), **T**rue **N**egative (Verdadero Negativo), **F**alse **P**ositive (Falso Positivo) y **F**alse **N**egative (Falso Negativo) de la predicción y generalmente se utiliza para obtener el rendimiento de un modelo de clasificación."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "id": "df710b46-4988-41f6-8fa0-711f6ffc0558",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[679,  18],\n",
       "       [ 47,  81]])"
      ]
     },
     "execution_count": 370,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "confusion_matrix(y_test, y_predicted)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8fc974d-d400-475d-a1bb-23d139c4cb86",
   "metadata": {},
   "source": [
    "Para mayor entendimiento, en la siguiente imagen se observan los parámetros de la matríz de confusión para este caso particular de un planteamiento de problema binario, como el presente, que clasifica si es **Spam o no**.\n",
    "\n",
    "<img src=\"img/matriz.png\" align=\"left\" style=\"width: 320px;\"/>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db2534e6-3a9b-4fa1-abea-02cb7466b50a",
   "metadata": {},
   "source": [
    "Considerando lo anterior y la definición dada de cada parámetro, efectuamos el cálculo de **El Accuracy** (Exactitud), que se refiere a lo cerca que está el resultado de una medición del valor verdadero, en otras palabras, la proporción de predicciones que el modelo clasificó correctamente."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 371,
   "id": "31a0111b-610a-4e84-90f3-e4a121b7f50b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9212121212121213"
      ]
     },
     "execution_count": 371,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "Y_pred_class= nb.predict(X_test_vect)\n",
    "metrics.accuracy_score(y_test, Y_pred_class)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "665d40d0-c80d-40ab-87ef-0dbb2c0b08b6",
   "metadata": {},
   "source": [
    "Como se observa en el resultado anterior, el 93.9% de las veces, el modelo es capaz de efectuar una buena predicción.\n",
    "Lo anterior se obtiene de la siguiente ecuación:\n",
    "\n",
    "$$\n",
    "\\text{Accuracy} = \\frac{TP+TN}{TP+TN+FP+FN}\n",
    "$$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbfa9089-7819-483d-b96b-9ac62d8d6ae9",
   "metadata": {},
   "source": [
    "### 1.2. Precision y recall:\n",
    "Ejecutamos una función para obtener el valor de **Precision** que representa el valor predictivo positivo y es la proporción de instancias relevantes entre las instancias recuperadas, en otras palabras, define qué proporción de identificaciones positivas fue realmente correcta."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 372,
   "id": "d9623518-77f2-4d95-b3db-7dfba5c696a6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8181818181818182"
      ]
     },
     "execution_count": 372,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, f1_score, classification_report, recall_score, confusion_matrix, precision_score\n",
    "precision_score(y_test, y_predicted) # con esta función obtenemos la precisión."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b9256b8-e830-4c22-9cd4-dd2e8f4f679c",
   "metadata": {},
   "source": [
    "Como se observa en el resultado anterior, el 88.7% de las veces, el modelo es capaz de efectuar una predicción positiva de manera correcta.\n",
    "Lo anterior se obtiene de la siguiente ecuación:\n",
    "\n",
    "$$\n",
    "\\text{Precision} = \\frac{TP}{TP+FP}\n",
    "$$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89ac8ed2-23f4-4b73-b237-dd844019a30c",
   "metadata": {},
   "source": [
    "En el caso del **Recall** (Sensibilidad), representa la proporción de la cantidad total de instancias pertinentes que se recuperaron realmente (True Positive Rate)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "id": "7c4c74dc-9745-47eb-ad5e-44f8d6d950ad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6328125"
      ]
     },
     "execution_count": 373,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_test, y_predicted) # con esta función obtenemos el recall."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fa0e4a3-bbc4-4384-832f-0893ce591460",
   "metadata": {},
   "source": [
    "Como se observa en el resultado anterior, el 68.8% representa la tasa de positividad verdadera (**T**rue **P**ositive **R**ate).\n",
    "Lo anterior se obtiene de la siguiente ecuación:\n",
    "\n",
    "$$\n",
    "\\text{Recall} = {TPR} = \\frac{TP}{TP+FN}\n",
    "$$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b871236f-9f5d-4396-a976-8faa44debc90",
   "metadata": {},
   "source": [
    "### 1.3. F1-Score:\n",
    "\n",
    "Esta métrica es una medida de la precisión de una prueba, es la media armónica de precisión y recuperación."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 374,
   "id": "7c7ccd8d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7136563876651982"
      ]
     },
     "execution_count": 374,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(y_test, y_predicted)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "322c3159-9c7f-464b-a189-220c32409f08",
   "metadata": {},
   "source": [
    "Puede tener una puntuación máxima de 1 (precisión y recuerdo perfectos) y una mínima de 0, representando la robustes del modelo, para este caso, se obtuvo 0.77.\n",
    "Lo anterior se obtiene de la siguiente ecuación:\n",
    "\n",
    "$$\n",
    "\\text{F1_Score} = \\frac{2*(Precision*Recall)}{Precision + Recall} = \\frac{2TP}{2TP+FP+FN}\n",
    "$$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe6a740a-dbbb-4240-ac1f-4f93b1a8f9e7",
   "metadata": {},
   "source": [
    "De igual manera, podemos obtener el reporte de clasificación, que permite obtener todas las métricas, representando las 2 clases (**Spam o no**)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 375,
   "id": "eb8e2138",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.97      0.95       697\n",
      "           1       0.82      0.63      0.71       128\n",
      "\n",
      "    accuracy                           0.92       825\n",
      "   macro avg       0.88      0.80      0.83       825\n",
      "weighted avg       0.92      0.92      0.92       825\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print (classification_report(y_test, y_predicted))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ff9e333a-36a0-4b15-8d48-6c5c2aab860d",
   "metadata": {},
   "source": [
    "### 1.4. Curva Precision V/S Recall:\n",
    "\n",
    "Para obtener la curva Precision v/s Recall, debemos generar una validación cruzada para obtener el puntaje, posteriormente podemos observar el `y_scores` y por último, plotear el registro de precisión y recall."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 376,
   "id": "f6890e53",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "# Generamos la función de decisión para generar el puntaje\n",
    "y_scores = cross_val_predict(nb, X_train_vect, y_train, cv=5, method='predict_proba') # le paso el clasificador,\n",
    "# datos de entrenamiento, y de entrenamiento y "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 377,
   "id": "608a1e3c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2475, 2)"
      ]
     },
     "execution_count": 377,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_scores.shape #cantidad de registros."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 378,
   "id": "424d04cc-7dd6-40a9-9b61-ba9979bee12c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([3.43460718e-03, 4.51601571e-06, 9.36620439e-01, ...,\n",
       "       1.21434226e-01, 9.97198123e-01, 1.13843696e-03])"
      ]
     },
     "execution_count": 378,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_scores[:,1] #sólo la columna de la derecha."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 379,
   "id": "74973503-1958-4531-8045-b411262a40ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_scores_proc=[]\n",
    "\n",
    "for row in y_scores:\n",
    "    y_scores_proc.append(row[1])\n",
    "    \n",
    "# Generamos las variables necesarias:\n",
    "precisions, recalls, thresholds = precision_recall_curve (y_train, y_scores_proc)\n",
    "\n",
    "def plot_precision_recall_vs_threshold(precisions, recalls, thresholds):\n",
    "    plt.plot(thresholds, precisions[:-1], 'b--', label='Precision')\n",
    "    plt.plot(thresholds, recalls[:-1], 'g-', label='Recall')\n",
    "    plt.legend(loc='center right', fontsize=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 380,
   "id": "12001982-4f85-4b97-8eb5-7a9c20dfe862",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plot_precision_recall_vs_threshold(precisions, recalls, thresholds)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "869bfc63-4877-45a8-a9ab-7609f36bad21",
   "metadata": {},
   "source": [
    "### 1.5. Curva ROC:\n",
    "\n",
    "Esta gráfica que enfrenta la tasa de falsos positivos (**T**rue **P**ositive **R**ate) (eje x) con la razón de falsos negativos (**F**alse **P**ositive **R**ate (eje y), permite medir falsa alarma en torno a un porcentaje de éxito.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 381,
   "id": "4fd22e36-287e-45d4-a05c-2efd3e68aebf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_train, y_scores_proc)\n",
    "\n",
    "def plot_roc_curve(fpr, tpr, label=None):\n",
    "    plt.plot(fpr, tpr, linewidth=3, label=label)\n",
    "    plt.plot([0, 1], [0, 1], 'k--')\n",
    "    plt.xlabel('False Positive Rate (FPR)', fontsize=14)\n",
    "    plt.ylabel('True Positive Rate (TPR)', fontsize=14)\n",
    "    \n",
    "plot_roc_curve(fpr, tpr)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9bf0d63e-c976-4bd0-9399-e88fa29182e9",
   "metadata": {},
   "source": [
    "Mientras más alejado de la linea punteada, es un mejor clasificador.\n",
    "\n",
    "De igual manera, se efectuará el cálculo del área debajo de la curva (AUC), que puede ser utilizado como resumen de la calidad del modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 382,
   "id": "140d8252-98f3-44d2-ab01-f1b6eeb5efbd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9181596694941685"
      ]
     },
     "execution_count": 382,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(y_train, y_scores_proc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "57b8ca40-bcab-4ac7-a3c5-baa8af1400f9",
   "metadata": {},
   "source": [
    "Como se observa y considerando que para efectuar comparación de clasificadores se utiliza la AUC, un clasificador perfecto tendrá un valor = 1, el modelo propuesto tiene un **0.91**."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ae277e77",
   "metadata": {},
   "source": [
    "## 2. Reducción de dimensionalidad y clustering\n",
    "\n",
    "En esta actividad vamos a trabajar con el _Dataset Iris_ para revisar conceptos de reducción de dimensionalidad y _clustering_.\n",
    "\n",
    "### 2.1. Reducción de dimensionalidad\\[1 pto\\]\n",
    "\n",
    "En esta parte vamos a hacer algunas pruebas clasificando elementos tras haber hecho reducción de dimensionalidad. Tomemos el código base de clases:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 383,
   "id": "248632a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "iris = load_iris()\n",
    "\n",
    "# Recordemos que el dataset original tiene 4 features\n",
    "X = iris['data']\n",
    "y = iris['target']\n",
    "\n",
    "# Recuerda que estas son las features\n",
    "iris['feature_names']\n",
    "\n",
    "# Ahora vamos a reducir la dimensionalidad\n",
    "# Queremos reducir de 4 a 2 dimensiones\n",
    "pca = PCA(n_components=2)\n",
    "\n",
    "# Guardamos el dataset pero ahora en su versión reducida en 2 dimensiones\n",
    "X_2D = pca.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 384,
   "id": "2e808bf9-28a5-4f7c-950b-73806845509a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['sepal length (cm)',\n",
       " 'sepal width (cm)',\n",
       " 'petal length (cm)',\n",
       " 'petal width (cm)']"
      ]
     },
     "execution_count": 384,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "iris['feature_names'] #Esto permite ver la cantidad de dimensiones que posee el dataset (4 en este caso)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 385,
   "id": "45686782",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "cmap_bold = ['#a72618', '#34892d', '#262ea7']\n",
    "\n",
    "plt.scatter(X [y==0, 0], X [y==0, 1], c=cmap_bold[0], label='Iris Setosa')\n",
    "plt.scatter(X [y==1, 0], X [y==1, 1], c=cmap_bold[1], label='Iris Versicolor')\n",
    "plt.scatter(X [y==2, 0], X [y==2, 1], c=cmap_bold[2], label='Iris Virginica')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b972f4a2",
   "metadata": {},
   "source": [
    "Ahora se pide que hagas un predictor del tipo de flor en base a los _datasets_ con menos dimensiones y el original. En concreto se pide que entrenes tres predictores con cada uno de los siguientes _datasets_:\n",
    "\n",
    "- El _dataset_ luego de aplicar PCA con una componente.\n",
    "- El _dataset_ luego de aplicar PCA con dos componentes.\n",
    "- El _dataset_ original.\n",
    "\n",
    "Para cada predictor se pide que midas el _accuracy_, la precisión y el _recall_ sobre un _dataset_ de prueba (la proporción entre entrenamiento - prueba la debes escoger tú). \n",
    "\n",
    "**Ojo**: la división de entrenamiento - prueba debe ser la misma para los tres casos, y en cada caso debes hacer `fit` de PCA solamente con el _dataset_ de entrenamiento, luego para el _dataset_ de prueba solamente debes usar  `transform`. Compara los resultados entre los predictores, y contesta la siguiente pregunta:\n",
    "\n",
    "> ¿Cómo se comporta el predictor a medida que bajan las dimensiones? ¿Crees que se pierde mucho al reducir dimensionalidad, en términos de _performance_ del clasificador?\n",
    "\n",
    "Para esta pregunta, puedes usar el clasificador que tú desees (por ejemplo, _random forest_)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ae19359-f31b-423f-9fb4-c34eedd4c11f",
   "metadata": {},
   "source": [
    "#### 2.1.1. El dataset con PCA de una componente utilizando el clasificador Random Forest.\n",
    "\n",
    "Para esto, se efectuará la reducción de las 4 dimensiones obtenidas inicialmente ('sepal length (cm)','sepal width (cm)','petal length (cm)','petal width (cm)') a un total de 1 y posteriormente se le aplicará un clasificador random forest, para luego, obtener sus métricas."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "id": "47978643",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "iris = load_iris()\n",
    "\n",
    "X = iris['data']\n",
    "y = iris['target']\n",
    "\n",
    "pca = PCA(n_components=1) # reducimos a uno la dimensionalidad.\n",
    "\n",
    "X_train_1, X_test_1, y_train_1, y_test_1 = train_test_split (X, y)\n",
    "\n",
    "X_train_1D = pca.fit_transform (X_train_1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b118a72d-e1b9-48f4-a1ce-e9827c29e231",
   "metadata": {},
   "source": [
    "Obtenido el `X_train_1D`, se implementa el clasificador Ramdon Forrest. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "id": "e048318a-f618-42dc-94d9-f6b8fb1df259",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.92316843])"
      ]
     },
     "execution_count": 299,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.explained_variance_ratio_ # varianzas por componente, este caso con 1."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95079cff-887e-44d5-90b5-839d21beb215",
   "metadata": {},
   "source": [
    "Con esto, comprendemos que la componente guarda un 92,3% de la varianza."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "id": "48bf715c-9366-41ea-b058-f58ffb962f72",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_1D = pca.transform(X_test_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "id": "8c7009ec-7f14-4a09-ba43-1c15d991eec1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2])"
      ]
     },
     "execution_count": 293,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rnd_clf1d = RandomForestClassifier(n_estimators=100, max_leaf_nodes=16, n_jobs=-1, random_state=42) # número de\n",
    "# estimadores y máximo de nodos en las hojas\n",
    "rnd_clf1d.fit(X_train_1D, y_train_1)\n",
    "rnd_clf1d.classes_ # con esto vemos las clases (Setosa, Versicolor, Virginica)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "id": "4a26b96a-6bf6-4d68-99e1-edf5f3155e45",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 1 0 2 1 2 2 2 1 0 0 1 2 2 2 1 2 2 0 0 2 2 2 2 2 0 2 2 0 0 1 2 2 1 1 1\n",
      " 2]\n"
     ]
    }
   ],
   "source": [
    "y_pred_rf_1D = rnd_clf1d.predict(X_test_1D)\n",
    "print(y_pred_rf_1D)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "id": "3bd3791f-a801-4c87-b705-018a03fb90eb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        10\n",
      "           1       0.89      0.57      0.70        14\n",
      "           2       0.68      0.93      0.79        14\n",
      "\n",
      "    accuracy                           0.82        38\n",
      "   macro avg       0.86      0.83      0.83        38\n",
      "weighted avg       0.84      0.82      0.81        38\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "print (classification_report(y_test_1, y_pred_rf_1D))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "id": "8d0435eb-c2a5-42b1-9a3b-b705bff7f74a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[10,  0,  0],\n",
       "       [ 0,  8,  6],\n",
       "       [ 0,  1, 13]])"
      ]
     },
     "execution_count": 296,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conf_mx_1D =confusion_matrix(y_test_1, y_pred_rf_1D)\n",
    "conf_mx_1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "id": "421a0676-cf36-4f07-b9fb-343aecacd464",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(conf_mx_1D, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "85fa9aa4-2200-49fd-8394-46cc4e923938",
   "metadata": {},
   "source": [
    "#### 2.1.2. El dataset con PCA de dos componentes utilizando el clasificador Random Forest:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 300,
   "id": "b2f23b54-00e4-45f5-9ae0-fb2da207ae6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "X = iris['data']\n",
    "y = iris['target']\n",
    "\n",
    "pca = PCA(n_components=2) # reducimos a dos la dimensionalidad.\n",
    "\n",
    "X_train_2, X_test_2, y_train_2, y_test_2 = train_test_split (X, y)\n",
    "\n",
    "X_train_2D = pca.fit_transform (X_train_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "id": "a95f4b37-4a47-4742-b385-b8e8ea97b0ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.93496715, 0.04370125])"
      ]
     },
     "execution_count": 302,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.explained_variance_ratio_ # varianzas por componente, este caso con 2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "832c3b58-c54c-4b16-baf4-0734bee752e1",
   "metadata": {},
   "source": [
    "La primera componente \"guarda\" un 93,5% de la varianza total, y la segunda componente guarda 4,4%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 308,
   "id": "032603c0-7ab5-4782-a374-9b8f368623d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2])"
      ]
     },
     "execution_count": 308,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_2D = pca.transform(X_test_2)\n",
    "\n",
    "rnd_clf2d = RandomForestClassifier(n_estimators=100, max_leaf_nodes=16, n_jobs=-1, random_state=42) # número de\n",
    "# estimadores y máximo de nodos en las hojas\n",
    "rnd_clf2d.fit(X_train_2D, y_train_2)\n",
    "rnd_clf2d.classes_ # con esto vemos las clases (Setosa, Versicolor, Virginica)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "id": "0cb335a6-19e6-452d-9d20-15991f6cca90",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2 2 1 0 1 0 1 1 0 1 2 1 2 1 0 1 0 2 1 2 2 2 0 2 1 0 0 2 0 1 1 1 2 2 1 0 1\n",
      " 0]\n"
     ]
    }
   ],
   "source": [
    "y_pred_rf_2D = rnd_clf2d.predict(X_test_2D)\n",
    "print(y_pred_rf_2D)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "id": "cb996749-00bc-42d8-ad41-ba5bb418bfe0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        11\n",
      "           1       0.93      0.93      0.93        15\n",
      "           2       0.92      0.92      0.92        12\n",
      "\n",
      "    accuracy                           0.95        38\n",
      "   macro avg       0.95      0.95      0.95        38\n",
      "weighted avg       0.95      0.95      0.95        38\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "print (classification_report(y_test_2, y_pred_rf_2D))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 311,
   "id": "03dca063-4ece-4c95-91ea-6214452e2a2e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[11,  0,  0],\n",
       "       [ 0, 14,  1],\n",
       "       [ 0,  1, 11]])"
      ]
     },
     "execution_count": 311,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conf_mx_2D =confusion_matrix(y_test_2, y_pred_rf_2D)\n",
    "conf_mx_2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "id": "cf8be08b-2f80-419d-8010-03421acc7d6f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(conf_mx_2D, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e52d40fe-d5c0-4f6a-877c-267ec3ce1f8c",
   "metadata": {},
   "source": [
    "#### 2.1.3. El dataset original utilizando el clasificador Random Forest:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "id": "5fc6753f-0254-41e7-9c13-dc0d12770622",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2])"
      ]
     },
     "execution_count": 273,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "X = iris['data']\n",
    "y = iris['target']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split (X, y)\n",
    "\n",
    "rnd_clf = RandomForestClassifier(n_estimators=100, max_leaf_nodes=16, n_jobs=-1, random_state=42) # número de\n",
    "# estimadores y máximo de nodos en las hojas\n",
    "\n",
    "rnd_clf.fit(X_train, y_train)\n",
    "rnd_clf.classes_ # con esto vemos las clases (Setosa, Versicolor, Virginica)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "id": "1d2aedef-7dd0-4351-922a-e2ab4c583808",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2 1 2 1 1 1 1 0 0 2 1 0 2 1 2 1 2 2 0 1 0 0 1 1 2 2 1 2 2 0 1 2 0 0 0 1 0\n",
      " 0]\n"
     ]
    }
   ],
   "source": [
    "y_pred_rf = rnd_clf.predict(X_test)\n",
    "print(y_pred_rf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "id": "70f125c0-3e5a-4705-996a-b8b5e84a87f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        12\n",
      "           1       1.00      0.93      0.97        15\n",
      "           2       0.92      1.00      0.96        11\n",
      "\n",
      "    accuracy                           0.97        38\n",
      "   macro avg       0.97      0.98      0.97        38\n",
      "weighted avg       0.98      0.97      0.97        38\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "print (classification_report(y_test, y_pred_rf))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "id": "c4a00c51-4ba9-4c8f-92cd-39c367c65c1a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[12,  0,  0],\n",
       "       [ 0, 14,  1],\n",
       "       [ 0,  0, 11]])"
      ]
     },
     "execution_count": 276,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conf_mx =confusion_matrix(y_test, y_pred_rf)\n",
    "conf_mx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "id": "50dd5c2b-9aad-4c32-ac0a-90135a6a9e08",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(conf_mx, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e05213e8-7424-4554-891a-c0dd07cad22c",
   "metadata": {},
   "source": [
    "#### 2.1.4. Conclusiones:\n",
    "\n",
    "**(Considerar que los porcentajes cambian, cada vez que se corre el código, debido a la asignación random que efectúa)**\n",
    "\n",
    " Si se efectúa una comparación de los 3 modelos obtenidos, se puede concluir que al reducir a sólo una dimensión, disminuye el Accuracy del modelo a un **89%**, comparado con el modelo original que tiene un **95%** de Accuracy.\n",
    " \n",
    "Por otra parte, si efectuamos una comparación del modelo original con el que aplica un PCA =  2, se obtiene un incremento en el Accuracy de **97%**.\n",
    " \n",
    "Por lo anterior, lo más conveniente al analizar cada modelo y aplicar técnicas de reducción de dimensionalidad, es efectuar una comparación de las métricas, de manera de no disminuir la eficiencia general del modelo."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf1fe526",
   "metadata": {},
   "source": [
    "### 2.2. Clustering\\[1 pto\\]\n",
    "\n",
    "En esta parte vamos a poner a prueba las técnicas de _clustering_ sobre el _dataset Iris_ para comprender que tan bien este algoritmo puede descubrir las etiquetas reales. En concreto se pide lo siguiente:\n",
    "\n",
    "- Ejecuta el algoritmo de KMeans sobre el _Dataset Iris_ con sus 4 features para descubrir **3 clusters**. Luego, compara los resultados del algoritmo con las etiquetas reales. ¿Qué tan bien crees que el algoritmo descubrió las etiquetas reales?\n",
    "- Luego repite el procedimiento sobre el _Dataset Iris_ luego de haber hecho PCA con 2 componentes. ¿Qué tanto cambia el resultado al haber reducido las dimensiones?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 321,
   "id": "9c6574d7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "iris = load_iris()\n",
    "\n",
    "# Recordemos que el dataset original tiene 4 features\n",
    "X = iris['data']\n",
    "y = iris['target']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "043c7677-7e94-4078-91d6-6401cc04ecbf",
   "metadata": {},
   "source": [
    "#### 2.2.1. Algoritmo de KMeans sobre el _Dataset Iris_ con sus 4 features para descubrir **3 clusters**:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 322,
   "id": "9a098bc7-808c-4a11-83c6-364fae6e3f6c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['sepal length (cm)',\n",
       " 'sepal width (cm)',\n",
       " 'petal length (cm)',\n",
       " 'petal width (cm)']"
      ]
     },
     "execution_count": 322,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Obtenemos las 4 features.\n",
    "iris['feature_names']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 329,
   "id": "582b6f81-04fa-4981-b459-f04a6782dd62",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importamos K-means\n",
    "from sklearn.cluster import KMeans\n",
    "\n",
    "k = 5 # elegimos 5 en forma aleatoria, tomando el ejemplo del otro notebook\n",
    "kmeans = KMeans(n_clusters=k, random_state=42)\n",
    "y_pred = kmeans.fit_predict(X) # como es aprendizaje no supervisado solo pasamos X, en general cuando hay un _\n",
    "# es una composición de dos funciones\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 330,
   "id": "a2168714-3c8b-4d7b-8417-ba7fbf275442",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 3, 3, 3, 2, 3, 3, 3, 2, 3, 2, 2, 3, 2, 3, 2, 3,\n",
       "       3, 2, 3, 2, 3, 2, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 3, 2, 3, 3, 3,\n",
       "       2, 2, 2, 3, 2, 2, 2, 2, 2, 3, 2, 2, 4, 3, 0, 4, 4, 0, 2, 0, 4, 0,\n",
       "       4, 4, 4, 3, 4, 4, 4, 0, 0, 3, 4, 3, 0, 3, 4, 0, 3, 3, 4, 0, 0, 0,\n",
       "       4, 3, 3, 0, 4, 4, 3, 4, 4, 4, 3, 4, 4, 4, 3, 4, 4, 3], dtype=int32)"
      ]
     },
     "execution_count": 330,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Vamos las etiquetas predecidas por el algoritmo K-means\n",
    "y_pred"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e6245b1-2eed-4a1a-bfd5-69b0f2ef487c",
   "metadata": {},
   "source": [
    "Se observa que aparecen 5 clusters (etiquetas 0, 1, 2, 3 y 4) de datos."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "id": "c7138ade-e535-4cb5-87c1-d5e757c755c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cmap_bold = ['#803260', '#3956b1', '#e614e3', '#64a0b9', '#6eab04']\n",
    "\n",
    "plt.figure(figsize=(6, 6))\n",
    "\n",
    "plt.scatter(X[y_pred==0, 0], X[y_pred==0, 1], c=cmap_bold[0], label='0') # filtro clasificados con cero\n",
    "plt.scatter(X[y_pred==1, 0], X[y_pred==1, 1], c=cmap_bold[1], label='1') # filtro clasificados con uno\n",
    "plt.scatter(X[y_pred==2, 0], X[y_pred==2, 1], c=cmap_bold[2], label='2') # filtro clasificados con dos\n",
    "plt.scatter(X[y_pred==3, 0], X[y_pred==3, 1], c=cmap_bold[3], label='3') # filtro clasificados con tres\n",
    "plt.scatter(X[y_pred==4, 0], X[y_pred==4, 1], c=cmap_bold[4], label='4') # filtro clasificados con cuatro\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 328,
   "id": "a316516c-23b1-41bc-b342-6bffef6e0d1e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "78.851441426146"
      ]
     },
     "execution_count": 328,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans.inertia_"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "20c05579-e15f-4074-b830-7267d3f163bb",
   "metadata": {},
   "source": [
    "Utilizamos el método del codo para determinar el número óptimo de clúster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "id": "d2a9ce30-fc5f-4456-acc8-1f3d8be76a02",
   "metadata": {},
   "outputs": [],
   "source": [
    "kmeans_per_k = [KMeans(n_clusters=k, random_state=42).fit(X)\n",
    "                for k in range(1, 10)]\n",
    "inertias = [model.inertia_ for model in kmeans_per_k]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "id": "931c245e-f638-4c28-b98c-393e324116b7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8, 3.5))\n",
    "plt.plot(range(1, 10), inertias, \"bo-\")\n",
    "plt.xlabel(\"$k$\", fontsize=14)\n",
    "plt.ylabel(\"Inertia\", fontsize=14)\n",
    "\n",
    "plt.axis([0, 10, 0, 700])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e0b9fcf-a096-43db-8f54-39926a1a4ce6",
   "metadata": {},
   "source": [
    "Como a través del método del codo obtuvimos que el número óptimo de clúster es = 3, ejecutamos nuevamente el algoritmo para k =3."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 386,
   "id": "95e3fc50-a774-4df7-9357-e7bb40bdd826",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "k = 3 # de acuerdo al enunciado\n",
    "kmeans = KMeans(n_clusters=k, random_state=42)\n",
    "y_pred = kmeans.fit_predict(X)\n",
    "cmap_bold = ['#803260', '#3956b1', '#e614e3']\n",
    "\n",
    "plt.scatter(X[y_pred==0, 0], X[y_pred==0, 1], c=cmap_bold[0], label='0') # filtro clasificados con cero\n",
    "plt.scatter(X[y_pred==1, 0], X[y_pred==1, 1], c=cmap_bold[1], label='1') # filtro clasificados con uno\n",
    "plt.scatter(X[y_pred==2, 0], X[y_pred==2, 1], c=cmap_bold[2], label='2') # filtro clasificados con dos\n",
    "\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "cmap_bold = ['#a72618', '#34892d', '#262ea7']\n",
    "\n",
    "plt.scatter(X [y==0, 0], X [y==0, 1], c=cmap_bold[0], label='Iris Setosa')\n",
    "plt.scatter(X [y==1, 0], X [y==1, 1], c=cmap_bold[1], label='Iris Versicolor')\n",
    "plt.scatter(X [y==2, 0], X [y==2, 1], c=cmap_bold[2], label='Iris Virginica')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7f75bafd-bdc3-4f92-b7d2-005fef88618e",
   "metadata": {},
   "source": [
    "Como se observa e la comparación de ambos gráficos, al aplicar un método de kmeans y descubrir el óptimo mediante la técnica del codo, observamos que el algoritmo tiene la capacidad de identificar de buena manera las etiquetas de clase (`Setosa`, `Versicolor`y `Virginica`).\n",
    "\n",
    "Lo anterior, no consideró una reducción de la dimensionalidad, como se efectuará en el siguiente etapa."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d26be57c-c705-471c-a4ae-473ba4532800",
   "metadata": {},
   "source": [
    "#### 2.2.2. Aprocedimiento sobre el _Dataset Iris_ luego de haber hecho PCA con 2 componentes:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 359,
   "id": "a5074104-8c4d-4f6e-8134-9fe85fe90ac2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "iris = load_iris()\n",
    "\n",
    "X = iris['data']\n",
    "y = iris['target']\n",
    "\n",
    "iris['feature_names']\n",
    "\n",
    "# Ahora vamos a reducir la dimensionalidad\n",
    "pca = PCA(n_components=2)\n",
    "\n",
    "# Guardamos el dataset pero ahora en su versión reducida en 2 dimensiones\n",
    "X_2D = pca.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 360,
   "id": "a6b1e4f5-7a9c-4df9-8ae6-a1c21b440681",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "\n",
    "k = 3 # de acuerdo al enunciado\n",
    "kmeans = KMeans(n_clusters=k, random_state=42)\n",
    "y_pred = kmeans.fit_predict(X_2D)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "id": "a4bc9437-0863-410b-be92-302480d7939f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAD8CAYAAAB0IB+mAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAAAwiElEQVR4nO3dfXxU5Znw8d+VSXCIiIBSCKCGdhUbIQlioCuVivgSi4kVC8LTVnm6an1YUZtahcoHlYctrnWphdaia6vVtQSt2Cam6ooFaVfFoItRQVR8pjUStqjlrSGSl/v5I5lhkpxz5u3MzJmZ6/v55FPmzMu5g/Q691znuq9bjDEopZTKfnnpHoBSSqnU0ICvlFI5QgO+UkrlCA34SimVIzTgK6VUjtCAr5RSOcKVgC8ivxSRv4rIWzbPnysi+0VkW8/PUjfOq5RSKnr5Ln3Ow8BPgUccXvNHY8wlLp1PKaVUjFyZ4RtjNgOfuvFZSimlksOtGX40/lFE3gB2AzcbY952evGJJ55oiouLUzIwpZTKFq+99trHxpjhVs+lKuC/DpxijDkkIl8Ffguc2vdFInItcC3AySefzNatW1M0PKWUyg4i8me751JSpWOMOWCMOdTz598DBSJyosXrHjDGnGWMOWv4cMsLlFJKqTilJOCLyEgRkZ4/T+457yepOLdSSqlurqR0RGQtcC5woog0A7cDBQDGmDXA14H/IyIdwGFgrtE2nUoplVKuBHxjzLwIz/+U7rJNpZRSaaIrbZVSKkdowFdKqSjVNwSYXlnH6eW1TK+so74hkO4hxSSVdfhKKZWx6hsCLFnWSFtbJwC7W1pZsqwRgKqZxWkcWfR0hq+UUlFYubopFOyD2to6Wbm6KU0jip0G/CQL1K+nbnoFtaePom56BYH69ekeklIqDi17WmM67kUa8JMoUL+exiU307q7GYyhdXczjUtu1qCvVAYqGlkY03Ev0oCfRE0rV9DZdrjXsc62wzStXJGmESml4lWzsBS/39frmN/vo2ZhaZpGFDu9aZtErS0fxXRcKeVdwRuzK1c30bKnlaKRhdQsLM2YG7agAT+pCotGd6dzLI4rpTJP1czijArwfWlKJ4lKaxbj8w/sdcznH0hpzeI0jUgplct0hp9ExVWzgO5cfmvLRxQWjaa0ZnHouFJKpZIG/CQrrpqlAV4p5Qma0lFKqRyhAV8ppXKEBnyllMoRGvCVUipHaMCPg/bHUUpF4sVWylqlE6Ngf5xgy4RgfxxAq3GUUoB3WynrDD9G2h9HKRWJV1spa8CPkfbHUUpF4tVWyhrwY2TXB0f74yilgrzaSlkDfoy0P45SKhKvtlLWm7Yx6tUfZ3cz4vP1yuHrjVullFdbKWvAj0MwqGu1jlLKjhdbKWtKJ05araOUyjQa8OOk1TpKqUyjAT9OWq2jlMo0GvDjpNU6SqlMozdt46S7WSmlMo0rAV9EfglcAvzVGDPe4nkBfgJ8FWgF5htjXnfj3Omku1kppTKJWymdh4FKh+cvBk7t+bkW+LlL51VKKRUlVwK+MWYz8KnDSy4FHjHdXgGGiEiRG+dWSikVnVTdtB0NfBj2uLnnWC8icq2IbBWRrXv37k3R0JRSKjd4qkrHGPOAMeYsY8xZw4cPT/dwlFIqq6Qq4H8EnBT2eEzPMaWUUimSqoBfB1wp3b4E7DfGtKTo3EoppXCvLHMtcC5woog0A7cDBQDGmDXA7+kuyXyf7rLM/+3GeZVSSkXPlYBvjJkX4XkD/LMb51JKKRUfT920VUoplTwa8NMsUL+euukV1J4+irrpFQTq16d7SEqpLKW9dNIoUL9eN1FRSqWMzvDTSDdRUUqlkgb8NNJNVJRSqaQBPw6B+vWsn1JC7bgiascV8eSUkrhy77qJilIqlTTgxyhQv55XF9/EkX1/Cx1r3/c3tiy6Keagr5uoKKVSSQN+jJpWrqCrvb3fcdPRHnPuvbhqFhXL76Fw1BgQoXDUGCqW36M3bJVSSaFVOjFyyq/Hk3vXTVSUSr36hgArVzfRsqeVopGF1CwspWpmcbqHlXQ6w4+RU35dc+9KeV99Q4AlyxrZ3dKKMbC7pZUlyxqpbwike2hJpwE/RqU1i8krKOh3XPILNPeuVAZYubqJtrbOXsfa2jpZubopTSNKHQ34MSqumsXkFfcyYMjQ0LGCIUOZcte9UadmdHWtUunTsqc1puPZRHP4cUgk766ra5VKr6KRhexu6R/ci0YWpmE0qaUz/BTT1bVKpVfNwlL8fl+vY36/j5qFpWkaUeroDD/FdHWtUukVrMbJxSodDfgpVlg0mtbdzZbHlXLTM9sbuG/zKv7nwB5GDB7Jgmk3cHHJzHQPyxOqZhbbBvhsLtnUlE6K6epalQrPbG/gh8/eyZ4DLRgMew608MNn7+SZ7Q3pHpqnZXvJpgb8FNPVtSoV7tu8iraOtl7H2jrauG/zqjSNKDNke8mmpnRcFKhfT9PKFbS2fERh0WhKaxZbBnJdXauS7X8O7InpuOqW7SWbOsN3SbDcsnV3MxgTKrfUGnuVDiMGj4zpuOpmV5qZLSWbGvBdouWWyksWTLsBf76/1zF/vp8F025I04gyQ7aXbGpKJ0590zdWlTeg5ZYqPYLVOFqlE5tsL9kUY0y6x2DprLPOMlu3bk33MCz1XS0LgAhY/F0WjhpD9cbGFI5OKZXLROQ1Y8xZVs9pSicOVukbq2Cv5ZZKKS/RgB+HaNI0BUOGarmlUilQ3xBgemUdp5fXMr2yLmtq5pNBA34colkVW1B4rAZ7pZLMSwulMuHCowE/DlarZfvSm7VKJZ9XFkp56cLjRAN+HHqtlrWhvXGUSj6vLJTyyoUnElcCvohUishOEXlfRBZZPD9fRPaKyLaen6vdOG86FVfNorRmMQVhG6EE5RXo7lfK+57Z3kDVmouYfHcZVWsuysg+O15ZKOWVC08kCQd8EfEBPwMuBkqAeSJSYvHSdcaY8p6fBxM9b7oFSzPb9/2t33MerXRVKiRbmqt5ZaGUVy48kbgxw58MvG+M+cAYcwSoBS514XM9zbI0s4fpaNcVtsrTsqW5WtXMYpYvrWBUUSEiMKqokOVLK1K+UMorF55I3FhpOxr4MOxxMzDF4nWXi8g04F3gu8aYDy1ekzEi3ZTVm7bKDcnqaZ8JzdWi7Uvv1Ns+VTJlhW6qWivUA2uNMZ+JyHeAXwHn9X2RiFwLXAtw8sknp2ho8XFqpxB8XqlEBNMuwZl4MO0CJBz0RwweyZ4DLZbHvSBY9RK8ERqsegE8F0SD+l54gmWaXroAuJHS+Qg4KezxmJ5jIcaYT4wxn/U8fBCYZPVBxpgHjDFnGWPOGj58uAtDSx6n0sxYVtgG6tdTN72C2tNHUTe9QrtrqpBkpl283lwtU6pe7Hi1TNONgN8InCoiY0VkADAXqAt/gYgUhT2sBna4cN60sivNFJ+P4svmRLXoSlsqKyfJTLtcXDKTH1TezsjBRQjCyMFF/KDyds80V8uUqhc7Xr1gJZzSMcZ0iMj1wHOAD/ilMeZtEVkGbDXG1AE3iEg10AF8CsxP9LxeEAzq4Y3UTGcngaceZ/ikyQCOG6I4tVTWVboq2WmXi0tmuhrg3dwLtmhkIbtb+gd3r1W92PHqBcuVHL4x5vfA7/scWxr258VAVham2wXt15YvoautLfRccPYORy8Udjd29Yavgu60S3gOHyKnXdK1cbnbOfeahaW9Pg+8WfVix6sXLF1pmyC74Ny+72+WF4Itt94QytcPOH6I5Xv1hq+C2NMu6aytdzuF4ZVyy3h5tUxT++EnqG56hWO1jhPJL0AEutrbQ8d8/oHaZVPFpWrNRZYpoJGDi6i/7rmknvv08lrLBYci8M62uUD8KR+3UkVuppy8eL4gp374uuNVgoq+MoNdax8Bjv5r9/kH4vP7OWKxCjec6Wgnf8hQ/IXHRtz4XKlI0llbHymFEW/Kx61UUTrKPL2wPqAvTekkIFC/nsBTjxMe7BGh+LI5nLlkecSOmgDt+/dRvbGRue/spnpjowZ7Fbd0blweKYURb8onmvdF05bY7nNuXbLF0+2M3aYBPwF2O1+1vPhC77JNEcTns/wMzdcrt6Sztj5Szj1S1Ypd0I7mfdHUu9t9TmeX8VSdfLJpSicKoQ3LdzcjPh+ms5PCUWMiblxeXDUrNGO32gdXt0BUdoLVNnsOtJAneXSZLkYOLnKsukn3xuVOKQynlI9TuiVSqsjpG0D4WOw+J9L7so0G/Aj6BmrT2f2PK9a2CsHA71SXrxT0b6nQZbqA6ForuF1b7xanMkunoP2VLxex9oldvZ4LTxVFW+9udf5o3pdtNOBH4NQV04rTrD18xq+UHauWCkHB1gpeDOpOnJqLff+2Vyzfs7ullafqA72OCXBZ1dFvEtHWu/c9f54InV39y4rSXSefbBrwI4hlEVThqDE6a1cJi1RVE3w+XYus4mWX8rEL2r486TcjN8CLf+ouPa1vCNB6uKPf++zq3cPP3zeN5PS+bKI3bW0E6tfz5JSSmHYz0Sob5YZIVTUjBo9MyiKrdO2AZVfhYzUDh+60SzBg79t3pNdzQ44viGqBVqYv7IqXLryyEKhfz5ZFN2E62iO/uIf4fFyxPb4FWEqF65vDD+fP9/ODyttDN3T7ineRldU5g+dKVWuGvumelaubLGf+o4q60y52z218tjrp400WNxZr6cKrGDWtXBFTsAf4/JxvJmk0KteEV9vYVenc/vQPLN9rlQ6KJvVj14r5rgfruXt7Z9JXi9qle+zSLnZ5/0y+6ZqKxWEa8C045u1F+MLcK/ng8f8IVewAtLz4AoH69ZrSUa5wqrZ5ZnsDIoLVt/O+6aBoN1GxulAc+X8lHNhyNnR2B9FUb0LidKPXbvafyTddoy0xTYTm8C04LYYqLBpNxR13MeVfV/VaSau97FUqBAN4sFQznNUiq2g3UbG6b/DZtunQOaD3e13o6R7NytigqpnFbHy2mne2zWXjs9WhwOfV5mSJSEVLZQ34FkprFiP5Bf2O5xUUhEounXrZK5UsdiWbeZJnmW+Ptr+O1Spd0zrY8r2JBCC3doLKxpuudt9O3PzWogHfQnHVLKbcdS8FQ4aGjg0YMpTJK+6NqZe9bl+oohVthYxdADfGWKaAou2vY9WKeeiJ1uHBKQBFmr1H6o3jxuw/U6XiW4vm8G1EWiRlt4l5MB3Ud4Wu1QYoSkFsm5XHugtWLJuo9L1vUH9SbLXq0dx0dEpbZOLG5W5yumfhFp3hx6noKzO6m32HCV9lqykfFa1YNiuPtUFaInvXxpo2iaazpVPawqv7wKZSsr+16Aw/DqG2yL2qJKTX5uW6faFyEl4qabBeC2OVvomnQVoi/XVi6ekezU1Hp5462Vhq6TVZF/BDnS2T2KDMur+OCZVmNq1cYbtCV9shK6eFVeHs0jRebZAWbV8bv98XCvhDji9gya2TsrbU0muyKqUTzJu37m4GY5JWKmk7e+85n10nTW2HrMC5OVpQqvrYu+WO5Y2WM/HwnL9VO4S2z46Wl2ZjqaXXZFXAT2bePLziRvKs/9rE57PtrFk4aozuVasA5+ZosebZveCO5Y2sfWJXvy+1Awf6euX8I+Xos7HU0muyKqWTrLy5XU/8cD7/QPs2yiJUb2wMfZb2xM9tdpU2qdhsPBkef/IDy+NHPuvqFayjyfF7cR/YbJJVM3y7/HiieXO7nvji84FIaPZeOGqM4/lTlXJS3pbOrQgTZVUnb9fVsu/xVCwsUs6yKuCX1izut3F4LHlzu4VSdt8QTFdXr83HI51fSzUVJFYqGS7V7YztVsn2qU4O8eX1fkJz9OmXVSmdRLYRdFooFWmRVbTn11JNFZRopU0si7XcYpeDHzjQx+HD/dOccy7/fK/HqVhYpJxpP/weddMrrIN6zy5WVhuQx3oT1ukcwRy/UtGoWnNRyu8DnF5ea7sfUF4edPXp5xZecqlSx6kfflaldBLhNPsurpp1NEcflrOP9WZroiknpYKibYrmJrtcuy9P+gV7gH372+NqjKaSRwN+j0g3fIurZlG9sTGUswdibozm1oVDZSY3c+7RNkVzU6xbEULutUbwOlcCvohUishOEXlfRBZZPH+MiKzreX6LiBS7cV43xTL7TqTapu+FQ4N9bnB7D9p0VPrY1ckHtxy0o60RvCPhm7Yi4gN+BlwANAONIlJnjNke9rJ/Av5mjPkHEZkL/CtwRaLndlMsN3ydqm00gCsrTg3S4rnJGk9PHTdEuxVhOLfLLt3Y9zVXuVGlMxl43xjzAYCI1AKXAuEB/1Lgjp4//wb4qYiI8dgd40gtkYO02kbFKhk5d6/01AkG2+V3v96rbQK4X3aZ6y2UE+VGSmc08GHY4+aeY5avMcZ0APuBE/p+kIhcKyJbRWTr3r17XRhacsS6wEs3QlGx5txTXWOfqKqZxWx5cRb3/PBLCbdGcNoERVsoJ8ZTdfjGmAeAB6C7LDPNw7FlV6bplO/XjVByWywbkaSjxt4tibZGiDSDT8W+r9nMjRn+R8BJYY/H9ByzfI2I5APHA5+4cO60iKXaRlfXZq9YZuGxrK6NZUOUbBNpBq/tGRLjxgy/EThVRMbSHdjnAv+rz2vqgKuAl4GvA3/wWv4+Vprvz23xzMKjzbnHku8P30glVTdukynSDN5pAxUVWcIz/J6c/PXAc8AO4HFjzNsiskxEqnte9gvgBBF5H6gB+pVueombOfdkNXRT6ZXMWXi0+X63Sz29INIMXlsoJ8aVOnxjzO+NMacZY75gjPmXnmNLjTF1PX9uM8bMNsb8gzFmcrCix4vc7mipq2uzU6RZeCI3XaOtsc/G1E80DdaSve9rNtOVtn1EyrkH6tfz5JQSascVUTuuiPVTShwvBrq6Njs5zcITnXlb5ftnjq/mvs2rel1AIl10nKpdvEpn8MmlzdP6qD19lPV+tCJ86Uc/ZcuimzAd7b2eyisoYPKKezWI5xCrfWn9+X5+UHk7921e5WpjM7tzHVPgZ//hfZbnufak+y1z3Ro8s582T4tCMG/vtPl408oV/YI9QFd7O00rV2i9fQ5xqrpxe5GVXerGGGOb+olU7ZKJs3+VOE/V4adL31r5voI591e+f73tZwRz/Vpvnzvsqm7stjCM1NjMruLG7kJxsO0Ad17yQ+7bvIoP3xhGe9MMDhwaxN3/2cnuFvtqF12tmrt0ho/9FoZwdPNxwHbzcrDewFzr7XNTPI3NnPL+TvcLLi6ZybUn3Y95fRYdhwYB3QHcZhMqikYW6mrVHKYBH4ea+LDNxxuX3Gy5eTl05/DtntN6+9wTzxaGdmmbOxpus/y2EH4BsQrgVonJYLVLslerarrIuzSlQ+QtDJ2+AUB3Dl98Psugr/X2uSnWxmZ2aZsu039nkZGDi3otsIo2UF9WVQxAngidFveq3Fitqukib9MZPpFr5aOZpVsFe623z15uNzeLduOSYKVP+MUk2kD9zPMfsmRZo+WGJbGuVrWbxUdzs3jKtCcZV1bLuLJapnxlvX4DSCEN+PSpledoPj5YeRPLLF18Pq23z3LJWOFqlfe3YvVNwGqxkpV9+45Y9qz35UlM5ZrBWfzullaMOTqLr28IOKaL6hsCLFq6hX37j1a67dt3hMW3v6pBP0W0Dj+MZbWOSE+ppmCdGe1DhLnv7E7WEJUHJGsD8fAqHRGxTedYnSN8U5A8EcdtB/sSgXe2zY369dMr6yyrgII7X8X6XPD5jc9W6+YmLtA6/ChZ5upDF8To/g+kOXtvSUZf+WRtIH5xyUzqr3uOV295gztm/ktMlT7h7Qb+dfkUy/YEQ44vsHxvrLl7p1m8U2sEp3sN4eWiVt8clDs04IdJtKJGc/be4mbqJfzCIWJd9OjmBuLxVPoE2bUnWHLrpIh9aqLh1ODMqTWC04VFy0VTQ6t0wthV6zgJVucUjhpjuweuSg+39pHt29rAKg2ajA3EE9nC0G4jktf+ey+PP/kBnV0GX55wWVXsG5ZEalFsd+6ahaUsWrqFjo7ef38FBXnULCzl+7e9Ynk+3dzEPTrDD2NVreNIhCu2NzN3ZwvVGxs12HuMW6kXqwsHQJ7khf43eCHxcmvi+oYAT9UHQvn9zi7DU/WBmFMm8TY4q5pZzF3LpvRKLQ0ZMoAVd052/Aagm5u4R2f4YYIBu2nlip6ZvvONWs3Xe1u8LQ76cqqR9+f7M2YrQqeUSayz/Hi3MnR6n25uknw6w++juGoW1RsbmbuzhS/d81MGDBlq+TrN13tfPC0OoP+N3sEDj7d8XXBmH87L/ei9vh+stkZOPp3hR9DZ1v+r/IAhQzlzyXJN4XhccJYdyxaAVlsX5ks+PvHRaY7OPAt8BbR39u+cColX67gpvNTTd+zCUL+dcF5KmSS6CbpylpMBP1C/vjtt0/IRhUWjbW+22rVUyC88VoN9hoj1xqdVvr7DdPR7XXtnOyJieQPXzWqdRPS9eOWXbqBjy0zoPJpD15RJbsm5gN93cZVTG2PdgDz3xDI7j6daJ5Wbjve9eA0Yux2Ajqbz6fz7IF3YlINyLuA7bWHYN+BHaqqmvMOtQGp3o9dJnuRhjIl4Xqt0UfhNXrcvBlYXrwFjt3PM2B28essbvY7rCtfckHMBP5ZZe2nN4n6tFvRmrfdECqSxWDDthn7bCUZijOkXQK1E2nTcrd8hKNoqJe1wmTtyrkrHbnZudVw3IM8MkQJpLKxWuFacPMXxPdHm7J3WBbj5OwRFW6WkK1xzR87N8GOdtRdXzdIA73Fu97YJv9Eb/PZgJ5YVtoMHHm+56fiIwSOT0p8n2iolr5drKvfk3AxfZ+3Zx2kLwETZrbIFYupv88z2Bv7edqjf8QJfAQum3ZC03yG8IduCaTdw3+ZV/RrJ6QrX3JFzM3zQWXu2scq7u9Xbxm6GLUhMrZDv27zKsrxzYEFh6IKRrN8BnO9z6ArX3JGTAV9ll3gWWEUr2e0ZDrYdAJL7OwQ/1+4eQfDCpVU62U8DvsoKiXSWdOLWt4doLhzJ+h0g8n0OXeGaGzTgK+XAjZn3M9sbONzef8V2Mloq23Hrm4rKbAkFfBEZBqwDioEAMMcY8zeL13UCb/Y8/IsxpjqR87oh2vYKSiUy8+6bOw8a7D+em89flLKumsm8z6EyR6JVOouAF4wxpwIv9Dy2ctgYU97zk9RgH6hfT930CmpPH0Xd9AoC9estX9O45ObuVbTGhNorWL1WKTvRbJ9oV+VTOKAwZcEeEttBq6/6hgDTK+s4vbyW6ZV1ugVhBkloE3MR2Qmca4xpEZEiYJMxZpzF6w4ZY/q36XMQzybmVpuQ+/wD+5Vd1k2vsG6ZMGoM1RsbYzqnyk1WM3d/vr9fEJ18dxnGYk8FQaJanes1fVflQndFj7Yx9o5kbmI+whgTTAzuAUbYvM4vIltF5BUR+VqC57Tl1CcnnDZFU4mKdmVsMtcIpIOuys1sEXP4IrIBsPrXeVv4A2OMERG7rwunGGM+EpHPA38QkTeNMbssznUtcC3AySefHHHwfUUbyLUpmopHeHMzq1k79K+GSTR3nsrumhC5iZquys1sEWf4xpjzjTHjLX5+B/xPTyqHnv/9q81nfNTzvx8Am4CJNq97wBhzljHmrOHDh8f8y0TbJ8dq71ptiqacBFM4ew602AZ76D9zTyR33vecwcVSydo3N5iu2d3SijFHm6iF5+h1VW5mSzSlUwdc1fPnq4Df9X2BiAwVkWN6/nwiMBXYnuB5LUUbyK3aKxRfNoemlSscb/aq3OXUYiHIbuYe3t6g/rrnop6hJ6OhGsAdyxspmbiOcWW1lExcxx3Lu+9bRZOuqVlYit/v6/UaXZWbORKtw78LeFxE/gn4MzAHQETOAq4zxlwNfBG4X0S66L7A3GWMSUrA77UJeYRyy/D2CrFsiqKyTzRpE6cmZoIkJd2SjIZqdyxvZO0TR7OpnV0m9DiadE0wvaOrcjNTQlU6yRRPlU68tGond0VbbVO15iLLhUsjBxfF1FMnFsk4Z8nEdXR29f//vC9PGDFiILtb+gf9UUWFbHw27UtnVJSSWaWTFbRqJ3dFmzaJtre8m2I9ZzT18VbBPnhc0zXZT1sroFU7uSzatEmym5tZieWc0e5a5csT2xm+pmuynwZ8dCvDXBZLj5lkNjezE+05nW64hgfsOZd/vlcOP/w4aBO1bKcpHXRTlFyWjlRNMkRbH3/Hkgrmzf4CvjwBumf282Z/gTuWVCR9jCr99KatynluLm5K9UKpoOmVdXrDVQHON201paOyVrTB1y5tEmvwdtpVKtlBX3etUtHQlI7KSomuUo3n/claKBWNqpnFLF9awaiiQkS6Z/ba0Ez1pSkdlZXsatiPHziEgQUDI87a46mBz7bOmCozaR2+yjl25Zb7D+/rNWtf+vRiKu4u7dfPPp5VrtnWGVNlH83hq6xkV25pp2++PZ4tAbN1V6n29naam5tpa3PuJaRSy+/3M2bMGAoKCqJ+jwZ8lZWsgm8kwXz7xSUz4wre6ViclQrNzc0cd9xxFBcXIyLpHo4CjDF88sknNDc3M3bs2KjfpwFfJVW6yhStgm/rkVYOtO13fF8wZRNv8E7H4iwnkfrbR6OtrU2DvceICCeccAJ79+6N7X1601YlS7SNydI5nr6S2Qwt1ay2IwQYcnwBS26dFHXg37FjB1/84heTMEKVKKv/NnrTVqVFOssUrYRvRmIlG/Lt4azaLQDs29/eb2MTrxs0yH5L7LPPPjvqz3n66aeZOHEiZWVllJSUcP/99zu+ftOmTbz00ktRf77XaUpHJU0y+rknKjzlkq50U6o4bTto1Wcn03R0dJCfnx91QG5vb+faa6/l1VdfZcyYMXz22WcEAgHH92zatIlBgwbFdFHxMp3hq6TxeplivDtRZYpI2w4max/aQP166qZXJGX3uE2bNnHOOedQXV1NSUkJcHT239LSwrRp0ygvL2f8+PH88Y9/7PXegwcP0tHRwQknnADAMcccw7hx4wDYu3cvl19+ORUVFVRUVPBf//VfBAIB1qxZw49//GPKy8v54x//SCAQ4LzzzqO0tJQZM2bwl7/8BYAnnniC8ePHU1ZWxrRp07r/HgIBzjnnHM4880zOPPNMT3xT0Bm+SppML1PM9G8AVu0WwiVjH9pU7B73+uuv89Zbb/WrTvn1r3/NRRddxG233UZnZyetrb0vaMOGDaO6uppTTjmFGTNmcMkllzBv3jzy8vK48cYb+e53v8uXv/xl/vKXv3DRRRexY8cOrrvuOgYNGsTNN3f/DlVVVVx11VVcddVV/PKXv+SGG27gt7/9LcuWLeO5555j9OjR7Nu3D4DPfe5zPP/88/j9ft577z3mzZtHuu9LasBXSZPJZYrp7IvjlmC6Zvndr7Nv35FezyWrz07TyhW92owDdLYdpmnlCtcC/uTJky1LESsqKvj2t79Ne3s7X/va1ygvL+/3mgcffJA333yTDRs2cM899/D888/z8MMPs2HDBrZvP7rz6oEDBzh06FC/97/88susX9/9jeVb3/oWt9xyCwBTp05l/vz5zJkzh1mzun/P9vZ2rr/+erZt24bP5+Pdd99149dPiAZ8lVReK1OMltMN50z6fYL97d0oz4xGKnaPO/bYYy2PT5s2jc2bN9PQ0MD8+fOpqanhyiuv7Pe6CRMmMGHCBL71rW8xduxYHn74Ybq6unjllVfw+/0WnxzZmjVr2LJlCw0NDUyaNInXXnuN1atXM2LECN544w26urri/mw3aQ5fKQtevOGciKqZxWx8tpp3ts1l47PVSbtZa7dLXCp2j/vzn//MiBEjuOaaa7j66qt5/fXXez1/6NAhNm3aFHq8bds2TjnlFAAuvPBCVq9e3es5gOOOO46DBw+Gjp999tnU1tYC8Nhjj3HOOecAsGvXLqZMmcKyZcsYPnw4H374Ifv376eoqIi8vDweffRROjutU2uppAFfKQtev+HsVaU1i/H5B/Y6lqrd4zZt2kRZWRkTJ05k3bp13Hjjjb2eN8Zw9913M27cOMrLy7n99tt5+OGHAVi1ahVbt26ltLSUkpIS1qxZA3Tn7J966qnQTdvVq1fz0EMPUVpayqOPPspPfvITAL7//e8zYcIExo8fz9lnn01ZWRkLFizgV7/6FWVlZbzzzju230xSSRdeKWXBa4vG0inWhVeB+vU0rVxBa8tHFBaNprRmse4elySxLrzSHL5SFjL5hnO6FVfN0gDvURrwlbKRqTeclbKjOXyllMoRGvD7SOYqQaWUSidN6YRJxSpBpZRKF53hh3FaJaiUUpkuoYAvIrNF5G0R6RIRyzKgntdVishOEXlfRBYlcs5kSsUqQaVU7Nxoj3znnXeyeHHv9QDbtm1LuNf/7t27+frXvx7Xe88999yU9tdJdIb/FjAL2Gz3AhHxAT8DLgZKgHkiUpLgeZMinasEc9Ez2xuoWnMRk+8u67eJuFKRdHR0AETdhXLevHmsW7eu17Ha2lrmzZsX0/n6GjVqFL/5zW+i+oxEJbpaN6GAb4zZYYzZGeFlk4H3jTEfGGOOALXApYmcN1nSuUow1wQXNu050ILBsOdAC0ufXsz5q6dp4M9wybyQJ9Ie+bTTTmPo0KFs2bIldOzxxx9n3rx57Nq1i8rKSiZNmsQ555zDO++8A8D8+fO57rrrmDJlCrfccgsvvvgi5eXllJeXM3HiRA4ePEggEGD8+PFAd0C++eabGT9+PKWlpaF2DS+88AITJ05kwoQJfPvb3+azzz7r97utXbs2tFr31ltvDR0fNGgQ3/ve9ygrK+Pll19O6O8vFTdtRwMfhj1uBqak4LwxC96Y1VWCyWfVnAxg/+F9GdeVUh2Vii6j8bZHhu5Zfm1tLVOmTOGVV15h2LBhnHrqqcyYMYM1a9Zw6qmnsmXLFhYsWMAf/vAHoHsT95deegmfz0dVVRU/+9nPmDp1KocOHerXEO2BBx4gEAiwbds28vPz+fTTT2lra2P+/Pm88MILnHbaaVx55ZX8/Oc/56abbgq9b/fu3dx666289tprDB06lAsvvJDf/va3fO1rX+Pvf/87U6ZM4d/+7d8S/ruLOMMXkQ0i8pbFj+uzdBG5VkS2isjWWDfndUtx1SyqNzYy953dVG9s1GCfJE5NyNK5DaJKTCq2tXRqj/zQQw9xxx138Oabb3Lcccf1e80VV1zBb37zG7q6ukLpnEOHDvHSSy8xe/ZsysvL+c53vkNLS0voPbNnz8bn8wHdbZBrampYtWoV+/btIz+/95x5w4YNfOc73wkdHzZsGDt37mTs2LGcdtppAFx11VVs3tw7C97Y2Mi5557L8OHDyc/P5xvf+EboNT6fj8svvzyBv7GjIgZ8Y8z5xpjxFj+/i/IcHwEnhT0e03PM6lwPGGPOMsacNXz48Cg/XmWiSE3IMrUrZa5LRZfRSO2RR48ezfz583nkkUf6veakk05i7NixvPjiizz55JNcccUVdHV1MWTIELZt2xb62bFjh+X5Fi1axIMPPsjhw4eZOnVqKPWTTH6/P3TBSVQqyjIbgVNFZKyIDADmAnUpOK/ysAXTbsCfb98fXLtSZqZ0dhmN1B45aN68eXz3u9/l85//PGPGjGHw4MGMHTuWJ554AujuqvnGG29YvnfXrl1MmDCBW2+9lYqKin4B/4ILLuD+++8P3eD99NNPGTduHIFAgPfffx+ARx99lK985Su93jd58mRefPFFPv74Yzo7O1m7dm2/17gh0bLMy0SkGfhHoEFEnus5PkpEfg9gjOkArgeeA3YAjxtj3k5s2CrTXVwykx9U3s5g//H9nsukbRBVb1YX8lT994zUHjlo9uzZvP32272qcx577DF+8YtfUFZWxhlnnMHvfmedwLj33ntDN2QLCgq4+OKLez1/9dVXc/LJJ1NaWkpZWRm//vWv8fv9PPTQQ8yePZsJEyaQl5fHdddd1+t9RUVF3HXXXUyfPp2ysjImTZrEpZe6X9ui7ZFV2mX63rHZLtb2yPrfM3W0PbLKONqVMrvof0/v0tYKSimVIzTgK6VUjtCAr5RSOUIDvlJK5QgN+EoplSM04CuVBPUNAaZX1nF6eS3TK+uobwike0gZLdntkeNtcfzVr36Vffv2Ob5m6dKlbNiwIebPTgatw1fKZfUNAZYsa6St7WgrW7/fx/KlFVTNLE7fwOIUax1+MgwaNIhDhw71OtbR0dGvl42Td999l8rKSj744IPQsUWLFlFYWMjSpUst3xPrOVIt1jp8neEr5bKVq5t6BXuAtrZOVq5uStOIUiuZ326S1R45vMXxww8/THV1Needdx4zZsygtbWVOXPmUFJSwmWXXcaUKVNCm5YUFxfz8ccfEwgE+OIXv8g111zDGWecwYUXXsjhw927582fPz/UL7+xsZGzzz6bsrIyJk+eHGqvfM4553DmmWdy5plnRt3fPx7evXQplaFa9vRvy+t0PJv0/Xazu6WVJcsaAVz7dpOM9siBQKDfOZqamhg2bBj33HMPQ4cOZfv27bz11luUl5dbjuu9995j7dq1/Pu//ztz5szhySef5Jvf/Gbo+SNHjnDFFVewbt06KioqOHDgAAMHDuRzn/sczz//PH6/n/fee4958+YlbRcsneEr5bKikYUxHc8mqfh243Z7ZCsXXHABw4YNA+BPf/oTc+fOBQj10bEyduzY0MVg0qRJ/S4iO3fupKioiIqKCgAGDx5Mfn4+7e3tXHPNNUyYMIHZs2ezffv2qP4e4qEBXymX1Swsxe/v3c7W7/dRs9A6UGSTVHy7cbs9cizncHLMMceE/uzz+Wy3ROzrxz/+MSNGjOCNN95g69atHDlyJOZzR0sDvvK8TNv7tmpmMcuXVjCqqBARGFVUmLE3bGOVzm838bZHjmTq1Kk8/vjjAGzfvp0333wzrvGNGzeOlpYWGhu7U1wHDx6ko6OD/fv3U1RURF5eHo8++mjC+9Y60Ry+8rRUbJmXDFUzi3MiwPdVs7DUskIpFd9uNm3axI9+9CMKCgoYNGiQ5Qwfutsj33DDDaH9ZiNZsGABV111FSUlJZx++umcccYZHH98/7bekQwYMIB169axcOFCDh8+zMCBA9mwYQMLFizg8ssv55FHHqGysjKubxfR0rJM5WlVay5iz4GWfsdHDi6i/rrn0jCi3BNrWWZ9Q4CVq5to2dNK0chCahaWZvTFr7Ozk/b2dvx+P7t27eL8889n586dDBgwIN1D0/bIKrukYss85a5s+3bT2trK9OnTaW9vxxjDfffd54lgHw8N+MrTRgweaTnD1y0QVaocd9xxSSuTTDW9aas8LZ1b5imVbXSGrzwteGNWt8xLL2MMIpLuYagw8dx/1YCvPE+3zEsvv9/PJ598wgknnKBB3yOMMXzyySf4/f7ILw6jAV8p5WjMmDE0Nzezd+/edA9FhfH7/VGtIwinAV8p5aigoMCylYHKPHrTVimlcoQGfKWUyhEa8JVSKkd4trWCiOwF/pzk05wIfJzkc7gpk8arY02eTBpvJo0VMmu8dmM9xRgz3OoNng34qSAiW+16TnhRJo1Xx5o8mTTeTBorZNZ44xmrpnSUUipHaMBXSqkckesB/4F0DyBGmTReHWvyZNJ4M2mskFnjjXmsOZ3DV0qpXJLrM3yllMoZOR/wReT/ikiTiGwTkf8UkVHpHpMdEfmRiLzTM96nRGRIusfkRERmi8jbItIlIp6sfBCRShHZKSLvi8iidI/HiYj8UkT+KiJvpXsskYjISSKyUUS29/wbuDHdY7IjIn4ReVVE3ugZ653pHlMkIuITkf8WkadjeV/OB3zgR8aYUmNMOfA0sDTN43HyPDDeGFMKvAssTvN4InkLmAVsTvdArIiID/gZcDFQAswTkZL0jsrRw0BlugcRpQ7ge8aYEuBLwD97+O/2M+A8Y0wZUA5UisiX0jukiG4EdsT6ppwP+MaYA2EPjwU8e1PDGPOfxpiOnoevALG1yksxY8wOY8zOdI/DwWTgfWPMB8aYI0AtcGmax2TLGLMZ+DTd44iGMabFGPN6z58P0h2cRqd3VNZMt0M9Dwt6fjwbB0RkDDATeDDW9+Z8wAcQkX8RkQ+Bb+DtGX64bwPPpHsQGW408GHY42Y8GpQymYgUAxOBLWkeiq2eFMk24K/A88YYz44VuBe4BeiK9Y05EfBFZIOIvGXxcymAMeY2Y8xJwGPA9V4ea89rbqP7K/Nj6RtpaCwRx6tyl4gMAp4EburzbdpTjDGdPWndMcBkERmf5iFZEpFLgL8aY16L5/050Q/fGHN+lC99DPg9cHsSh+Mo0lhFZD5wCTDDeKCmNoa/Wy/6CDgp7PGYnmPKBSJSQHewf8wYsz7d44mGMWafiGyk+16JF2+OTwWqReSrgB8YLCL/YYz5ZjRvzokZvhMROTXs4aXAO+kaSyQiUkn3V7lqY0xruseTBRqBU0VkrIgMAOYCdWkeU1aQ7r0QfwHsMMasTPd4nIjI8GDFm4gMBC7Ao3HAGLPYGDPGGFNM97/XP0Qb7EEDPsBdPSmIJuBCuu9+e9VPgeOA53vKSNeke0BOROQyEWkG/hFoEJHn0j2mcD03wK8HnqP7puLjxpi30zsqeyKyFngZGCcizSLyT+kek4OpwLeA83r+rW7rmZV6URGwsScGNNKdw4+p3DFT6EpbpZTKETrDV0qpHKEBXymlcoQGfKWUyhEa8JVSKkdowFdKqRyhAV8ppXKEBnyllMoRGvCVUipH/H/ymeVjrO1IrwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "cmap_bold = ['#803260', '#3956b1', '#e614e3', '#64a0b9', '#6eab04']\n",
    "\n",
    "plt.scatter(X_2D[y_pred==0, 0], X_2D[y_pred==0, 1], c=cmap_bold[0], label='0') # filtro clasificados con cero\n",
    "plt.scatter(X_2D[y_pred==1, 0], X_2D[y_pred==1, 1], c=cmap_bold[1], label='1') # filtro clasificados con uno\n",
    "plt.scatter(X_2D[y_pred==2, 0], X_2D[y_pred==2, 1], c=cmap_bold[2], label='2') # filtro clasificados con dos\n",
    "\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "cmap_bold = ['#a72618', '#34892d', '#262ea7']\n",
    "\n",
    "plt.scatter(X_2D[y==0, 0], X_2D[y==0, 1], c=cmap_bold[0], label='Iris Setosa')\n",
    "plt.scatter(X_2D[y==1, 0], X_2D[y==1, 1], c=cmap_bold[1], label='Iris Versicolor')\n",
    "plt.scatter(X_2D[y==2, 0], X_2D[y==2, 1], c=cmap_bold[2], label='Iris Virginica')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49096382-6ead-40ba-843b-ad146148744a",
   "metadata": {},
   "source": [
    "Como se observa en la figura, se obtienen similares resultados al aplicar clustering para reconocimiento de etiquetas de clase (`Setosa`, `Versicolor`y `Virginica`), tras aplicar técnicas de reducción de dimensionalidad.\n",
    "\n",
    "A pesar de haber reducido las features (dimensiones), el algoritmo Kmeans de igual manera reconoce etiquetas."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5815e0a6",
   "metadata": {},
   "source": [
    "## 3. Random Forest\n",
    "\n",
    "En esta parte vas a entrenar un predictor para el _Dataset Fashion MNist_, que es similar al famoso _dataset_ MNist, pero en este las imágenes corresponden a prendas. La idea es que al ver una imagen, podamos predecir el tipo de prenda. Primero vamos a cargar el _dataset_."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 363,
   "id": "c2c92f39",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pip install tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 396,
   "id": "efc403b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 396,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tensorflow import keras\n",
    "\n",
    "fashion_mnist = keras.datasets.fashion_mnist\n",
    "\n",
    "# Cargamos el dataset, notemos que ya viene dividido\n",
    "(X_train_full, y_train_full), (X_test, y_test) = fashion_mnist.load_data()\n",
    "X_train_full.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5ce7520-aa14-4e96-8830-72400b285f48",
   "metadata": {},
   "source": [
    "De acuerdo a lo detallado anteriormente, se definen las siguientes variables para su clasificación: `T-shirt/top`, `Trouser`, `Pullover`, `Dress`, `Coat`,\n",
    "               `Sandal`, `Shirt`, `Sneaker`, `Bag` y `Ankle boot`."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "903c3ce8",
   "metadata": {},
   "source": [
    "Las clases son las siguientes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 408,
   "id": "7a278118",
   "metadata": {},
   "outputs": [],
   "source": [
    "class_names = [\"T-shirt/top\", \"Trouser\", \"Pullover\", \"Dress\", \"Coat\",\n",
    "               \"Sandal\", \"Shirt\", \"Sneaker\", \"Bag\", \"Ankle boot\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "919adffc",
   "metadata": {},
   "source": [
    "Vamos a graficar parte del _dataset_."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 402,
   "id": "51bc8a6b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x345.6 with 40 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "n_rows = 4\n",
    "n_cols = 10\n",
    "plt.figure(figsize=(n_cols * 1.2, n_rows * 1.2))\n",
    "for row in range(n_rows):\n",
    "    for col in range(n_cols):\n",
    "        index = n_cols * row + col\n",
    "        plt.subplot(n_rows, n_cols, index + 1)\n",
    "        plt.imshow(X_train_full[index], cmap=\"binary\", interpolation=\"nearest\")\n",
    "        plt.axis('off')\n",
    "        plt.title(class_names[y_train_full[index]], fontsize=12)\n",
    "plt.subplots_adjust(wspace=0.2, hspace=0.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "642ef666",
   "metadata": {},
   "source": [
    "Tenemos que cambiar la dimensión de las imagenes de $28\\times28$ a $784$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 403,
   "id": "c007e914",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_full = X_train_full.reshape(X_train_full.shape[0], 784)\n",
    "X_test = X_test.reshape(X_test.shape[0], 784)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34659a47",
   "metadata": {},
   "source": [
    "### 3.1. Visualización con T-SNE \\[1 pto\\]\n",
    "\n",
    "Lo primero que debes hacer es visualizar el _dataset_ en dos dimensiones utilizanddo T-SNE (esto te puede servir para entender el _dataset_ a un nivel más profundo. ¿Puedes distinguir los _clusters_ claramente? ¿Qué puedes decir de los elementos de distintas clases que quedan cerca entre sí? Recuerda intentar con un número limitado de imágenes para que tome menos tiempo (ej. 2000 primeras observaciones)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 409,
   "id": "50e7c789",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "\n",
    "cmap_bold = ['#803260', '#3956b1', '#e614e3', '#64a0b9', '#6eab04', \n",
    "             '#467b5c', '#17fa19', '#926409', '#becdc7', '#f8acaf']\n",
    "tsne = TSNE(n_components=2)\n",
    "\n",
    "X_train_tsne = X_train_full [:2000]\n",
    "y_train_tsne = y_train_full [:2000]\n",
    "X_2D = tsne.fit_transform (X_train_tsne)\n",
    "\n",
    "plt.figure (figsize =(9,9))\n",
    "\n",
    "plt.scatter (X_2D[y_train_tsne==0, 0], X_2D[y_train_tsne==0, 1], c=cmap_bold[0], label=class_names[0])\n",
    "plt.scatter (X_2D[y_train_tsne==1, 0], X_2D[y_train_tsne==1, 1], c=cmap_bold[1], label=class_names[1])\n",
    "plt.scatter (X_2D[y_train_tsne==2, 0], X_2D[y_train_tsne==2, 1], c=cmap_bold[2], label=class_names[2])\n",
    "plt.scatter (X_2D[y_train_tsne==3, 0], X_2D[y_train_tsne==3, 1], c=cmap_bold[3], label=class_names[3])\n",
    "plt.scatter (X_2D[y_train_tsne==4, 0], X_2D[y_train_tsne==4, 1], c=cmap_bold[4], label=class_names[4])\n",
    "plt.scatter (X_2D[y_train_tsne==5, 0], X_2D[y_train_tsne==5, 1], c=cmap_bold[5], label=class_names[5])\n",
    "plt.scatter (X_2D[y_train_tsne==6, 0], X_2D[y_train_tsne==6, 1], c=cmap_bold[6], label=class_names[6])\n",
    "plt.scatter (X_2D[y_train_tsne==7, 0], X_2D[y_train_tsne==7, 1], c=cmap_bold[7], label=class_names[7])\n",
    "plt.scatter (X_2D[y_train_tsne==8, 0], X_2D[y_train_tsne==8, 1], c=cmap_bold[8], label=class_names[8])\n",
    "plt.scatter (X_2D[y_train_tsne==9, 0], X_2D[y_train_tsne==9, 1], c=cmap_bold[9], label=class_names[9])\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "98a3d1a3",
   "metadata": {},
   "source": [
    "Como se observa en el gráfico, al efectuar la separación de etiquetas por reconococimiento, el T-Sne tiende a confundir `Pullover` con `Shirt` y con `Coat`, que parecen similares fotografías en la muestra descrita anteriormente.\n",
    "\n",
    "Por otra parte, elementos como `Trouser`, `Bag`, `T-Shirt/Top`, `Sandal` y `Ankle boot`, tienden a ser correctamente identificados debido a que sus características físicas permiten distinguirse correctamente de los otros."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "364be07e",
   "metadata": {},
   "source": [
    "### 3.2. Entrenando el modelo \\[1 pto\\]\n",
    "\n",
    "Ahora debes entrenar un modelo de `Random Forest` para predecir el tipo de prenda del _dataset_ anterior. Debes justificar la elección de los hiperparámetros (por ejemplo, prueba varias combinaciones y decide cuál es la mejor), y además, debes presentar la evaluación del modelo utilizando los datos de prueba que cargamos anteriormente (además de accuracy, puedes presentar una matriz de confusión). ¿Qué opinas del desempeño del predictor? ¿Crees que se puede mejorar?"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bdba4404-920d-49fa-b33c-93fea91bdda4",
   "metadata": {},
   "source": [
    "#### 3.2.1. Modelo con hiperparámetros (64 estimadores, hojas máx.: 16):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 476,
   "id": "c711a345",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_leaf_nodes=16, n_estimators=64, n_jobs=-1,\n",
       "                       random_state=42)"
      ]
     },
     "execution_count": 476,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rnd_clf16 = RandomForestClassifier(n_estimators=64, max_leaf_nodes=16, n_jobs=-1, random_state=42) # modificar el número de hiperparámetros.\n",
    "rnd_clf16.fit(X_train_full, y_train_full)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 477,
   "id": "5ddc8c8d-f5e5-4526-b5bb-42e51ab269e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "class_names = [\"T-shirt/top\", \"Trouser\", \"Pullover\", \"Dress\", \"Coat\",\n",
    "               \"Sandal\", \"Shirt\", \"Sneaker\", \"Bag\", \"Ankle boot\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 478,
   "id": "db1f9836-ec71-4bf8-ac8c-04fe33dfe187",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=uint8)"
      ]
     },
     "execution_count": 478,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_clf16.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 479,
   "id": "2cb72c91",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.78      0.76      1000\n",
      "           1       0.96      0.85      0.90      1000\n",
      "           2       0.57      0.67      0.62      1000\n",
      "           3       0.64      0.85      0.73      1000\n",
      "           4       0.51      0.76      0.61      1000\n",
      "           5       0.96      0.83      0.89      1000\n",
      "           6       0.65      0.02      0.04      1000\n",
      "           7       0.84      0.89      0.86      1000\n",
      "           8       0.87      0.92      0.89      1000\n",
      "           9       0.86      0.93      0.89      1000\n",
      "\n",
      "    accuracy                           0.75     10000\n",
      "   macro avg       0.76      0.75      0.72     10000\n",
      "weighted avg       0.76      0.75      0.72     10000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, classification_report\n",
    "\n",
    "y_pred_rf = rnd_clf16.predict(X_test)\n",
    "\n",
    "print(classification_report(y_test, y_pred_rf))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a8fdee6f-79a3-4fb0-84fb-5dcdc8643c18",
   "metadata": {},
   "source": [
    "#### 3.2.2. Modelo con hiperparámetros (100 estimadores, hojas máx.: 100):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 480,
   "id": "aad498f3-3229-4971-a24f-abb78eead7fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_leaf_nodes=100, n_jobs=-1, random_state=42)"
      ]
     },
     "execution_count": 480,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rnd_clf100 = RandomForestClassifier(n_estimators=100, max_leaf_nodes=100, n_jobs=-1, random_state=42) # modificar el número de hiperparámetros.\n",
    "rnd_clf100.fit(X_train_full, y_train_full)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 481,
   "id": "f3b1526b-880f-4ab8-9459-ff812c28c8ed",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.83      0.79      1000\n",
      "           1       0.99      0.93      0.96      1000\n",
      "           2       0.68      0.72      0.70      1000\n",
      "           3       0.80      0.88      0.84      1000\n",
      "           4       0.67      0.78      0.72      1000\n",
      "           5       0.97      0.92      0.94      1000\n",
      "           6       0.70      0.40      0.51      1000\n",
      "           7       0.90      0.90      0.90      1000\n",
      "           8       0.92      0.96      0.94      1000\n",
      "           9       0.90      0.94      0.92      1000\n",
      "\n",
      "    accuracy                           0.83     10000\n",
      "   macro avg       0.83      0.83      0.82     10000\n",
      "weighted avg       0.83      0.83      0.82     10000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, classification_report\n",
    "\n",
    "y_pred_rf = rnd_clf100.predict(X_test)\n",
    "\n",
    "print(classification_report(y_test, y_pred_rf))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "468bc695-1b98-4402-a1ef-cbcd41500c4d",
   "metadata": {},
   "source": [
    "#### 3.2.3. Modelo con hiperparámetros (200 estimadores, hojas máx.: 1000):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 486,
   "id": "666b89fd-0033-4697-8a16-0f2977775ff1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_leaf_nodes=1000, n_estimators=200, n_jobs=-1,\n",
       "                       random_state=42)"
      ]
     },
     "execution_count": 486,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rnd_clf_1000 = RandomForestClassifier(n_estimators=200, max_leaf_nodes=1000, n_jobs=-1, random_state=42) # modificar el número de hiperparámetros.\n",
    "rnd_clf_1000.fit(X_train_full, y_train_full)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 487,
   "id": "cfcf41e9-3580-4b22-9dab-6025e9535968",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.85      0.83      1000\n",
      "           1       0.99      0.95      0.97      1000\n",
      "           2       0.75      0.78      0.77      1000\n",
      "           3       0.85      0.90      0.88      1000\n",
      "           4       0.74      0.81      0.77      1000\n",
      "           5       0.97      0.95      0.96      1000\n",
      "           6       0.71      0.55      0.62      1000\n",
      "           7       0.92      0.94      0.93      1000\n",
      "           8       0.95      0.97      0.96      1000\n",
      "           9       0.94      0.95      0.95      1000\n",
      "\n",
      "    accuracy                           0.87     10000\n",
      "   macro avg       0.86      0.87      0.86     10000\n",
      "weighted avg       0.86      0.87      0.86     10000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, classification_report\n",
    "\n",
    "y_pred_rf = rnd_clf_1000.predict(X_test)\n",
    "\n",
    "print(classification_report(y_test, y_pred_rf))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39384389-b7d4-4867-90bb-db25132d9847",
   "metadata": {},
   "source": [
    "De acuerdo a los resultados obtenidos, se puede señalar lo siguiente: el número máximo de hojas (max_leaf_nodes) afecta directamente al accuracy obtenido en cada reporte, pasando inicialmente de 16 hojas con un **75%** a un **87%**, con el aumento.\n",
    "\n",
    "Por lo anterior, se decidió mantener el último seteo de hiperparámetros propuesto."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 488,
   "id": "b774d15a-c8fb-4b05-95b3-2bd040ebcc0b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.8701 , 0.8759 , 0.87455])"
      ]
     },
     "execution_count": 488,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "cross_val_score(rnd_clf_1000, X_train_full, y_train_full, cv=3, scoring='accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 489,
   "id": "55a7f10a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[849,   1,  11,  40,   4,   1,  79,   0,  15,   0],\n",
       "       [  2, 948,   7,  30,   3,   0,   8,   0,   2,   0],\n",
       "       [  9,   0, 781,  10, 140,   0,  54,   0,   6,   0],\n",
       "       [ 19,   3,  12, 902,  34,   0,  28,   0,   2,   0],\n",
       "       [  0,   1,  94,  39, 813,   0,  50,   0,   3,   0],\n",
       "       [  0,   0,   0,   1,   0, 949,   0,  34,   2,  14],\n",
       "       [161,   2, 128,  34, 101,   0, 552,   0,  22,   0],\n",
       "       [  0,   0,   0,   0,   0,  18,   0, 940,   0,  42],\n",
       "       [  0,   2,   6,   2,   5,   2,   8,   4, 969,   2],\n",
       "       [  0,   0,   0,   0,   0,   8,   1,  41,   2, 948]])"
      ]
     },
     "execution_count": 489,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_pred = cross_val_predict(rnd_clf_1000, X_train_full, y_train_full, cv=3)\n",
    "conf_mx = confusion_matrix(y_test, y_pred_rf)\n",
    "conf_mx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 490,
   "id": "b9f6fb85-0e61-4866-b4f8-1d950ab34c72",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x864 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
    "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']\n",
    "\n",
    "fig = plt.figure(figsize =(12,12))\n",
    "ax = fig.add_subplot(111)\n",
    "cax = ax.matshow(conf_mx, cmap=plt.cm.gray)\n",
    "fig.colorbar(cax)\n",
    "\n",
    "xaxis = np.arange(len(class_names))\n",
    "ax.set_xticks(xaxis)\n",
    "ax.set_yticks(xaxis)\n",
    "ax.set_xticklabels(class_names)\n",
    "ax.set_yticklabels(class_names)\n",
    "\n",
    "plt.title ('Matriz de Confusión', fontsize=20)\n",
    "plt.ylabel ('Valores de Predicción', fontsize=16)\n",
    "plt.xlabel ('Valores verdaderos', fontsize=16)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19bea8f0-d3e2-4f11-afe1-ebab598f7c54",
   "metadata": {},
   "source": [
    "Mientras más instancias caen en un determinado casillero, más blanco el casillero. El casillero con el menor valor se lleva el color negro, y el de mayor valor el color blanco."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 491,
   "id": "fb49b78f-fa48-4418-99b9-9f12a8ab0993",
   "metadata": {},
   "outputs": [],
   "source": [
    "row_sums = conf_mx.sum(axis=1, keepdims=True)\n",
    "norm_conf_mx = conf_mx / row_sums"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "812c3b51",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 492,
   "id": "f87efc8d-45cf-454e-bafe-aa67f032c2ef",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x864 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "np.fill_diagonal(norm_conf_mx, 0)\n",
    "\n",
    "fig = plt.figure(figsize =(12,12))\n",
    "ax = fig.add_subplot(111)\n",
    "cax = ax.matshow(norm_conf_mx, cmap=plt.cm.gray)\n",
    "fig.colorbar(cax)\n",
    "\n",
    "xaxis = np.arange(len(class_names))\n",
    "ax.set_xticks(xaxis)\n",
    "ax.set_yticks(xaxis)\n",
    "ax.set_xticklabels(class_names)\n",
    "ax.set_yticklabels(class_names)\n",
    "plt.title ('Matriz de Confusión Normalizada', fontsize=20)\n",
    "plt.ylabel ('Valores de Predicción', fontsize=16)\n",
    "plt.xlabel ('Valores verdaderos', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6326f40c-cbdd-43f5-9055-cb06d5a1f41e",
   "metadata": {},
   "source": [
    "Al observar la matriz de confusión normalizada, podemos determinar donde existen las desinformaciones, esto es errores de clasificación.\n",
    "Al igual que en la utilización del T-Sne, se producen distorsiones al clasificar `Pull Over`, `Dress`, `Coat` y en menor medida `Shirt`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5f80cfaf-dd43-442b-a92c-689cdb3e69a5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: nbconvert[webpdf] in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (6.1.0)\n",
      "Requirement already satisfied: pygments>=2.4.1 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (2.10.0)\n",
      "Requirement already satisfied: nbclient<0.6.0,>=0.5.0 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (0.5.4)\n",
      "Requirement already satisfied: entrypoints>=0.2.2 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (0.3)\n",
      "Requirement already satisfied: traitlets>=5.0 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (5.0.5)\n",
      "Requirement already satisfied: pandocfilters>=1.4.1 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (1.4.3)\n",
      "Requirement already satisfied: jinja2>=2.4 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (3.0.1)\n",
      "Requirement already satisfied: jupyterlab-pygments in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (0.1.2)\n",
      "Requirement already satisfied: nbformat>=4.4 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (5.1.3)\n",
      "Requirement already satisfied: mistune<2,>=0.8.1 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (0.8.4)\n",
      "Requirement already satisfied: bleach in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (4.0.0)\n",
      "Requirement already satisfied: defusedxml in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (0.7.1)\n",
      "Requirement already satisfied: jupyter-core in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (4.7.1)\n",
      "Requirement already satisfied: testpath in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbconvert[webpdf]) (0.5.0)\n",
      "Collecting pyppeteer==0.2.2\n",
      "  Downloading pyppeteer-0.2.2-py3-none-any.whl (145 kB)\n",
      "\u001b[K     |████████████████████████████████| 145 kB 6.9 MB/s eta 0:00:01\n",
      "\u001b[?25hCollecting websockets<9.0,>=8.1\n",
      "  Downloading websockets-8.1.tar.gz (58 kB)\n",
      "\u001b[K     |████████████████████████████████| 58 kB 9.1 MB/s  eta 0:00:01\n",
      "\u001b[?25hRequirement already satisfied: urllib3<2.0.0,>=1.25.8 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from pyppeteer==0.2.2->nbconvert[webpdf]) (1.26.6)\n",
      "Collecting tqdm<5.0.0,>=4.42.1\n",
      "  Downloading tqdm-4.62.2-py2.py3-none-any.whl (76 kB)\n",
      "\u001b[K     |████████████████████████████████| 76 kB 6.5 MB/s  eta 0:00:01\n",
      "\u001b[?25hCollecting pyee<8.0.0,>=7.0.1\n",
      "  Downloading pyee-7.0.4-py2.py3-none-any.whl (12 kB)\n",
      "Collecting appdirs<2.0.0,>=1.4.3\n",
      "  Downloading appdirs-1.4.4-py2.py3-none-any.whl (9.6 kB)\n",
      "Requirement already satisfied: MarkupSafe>=2.0 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jinja2>=2.4->nbconvert[webpdf]) (2.0.1)\n",
      "Requirement already satisfied: nest-asyncio in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbclient<0.6.0,>=0.5.0->nbconvert[webpdf]) (1.5.1)\n",
      "Requirement already satisfied: jupyter-client>=6.1.5 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbclient<0.6.0,>=0.5.0->nbconvert[webpdf]) (7.0.1)\n",
      "Requirement already satisfied: pyzmq>=13 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jupyter-client>=6.1.5->nbclient<0.6.0,>=0.5.0->nbconvert[webpdf]) (22.2.1)\n",
      "Requirement already satisfied: python-dateutil>=2.1 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jupyter-client>=6.1.5->nbclient<0.6.0,>=0.5.0->nbconvert[webpdf]) (2.8.2)\n",
      "Requirement already satisfied: tornado>=4.1 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jupyter-client>=6.1.5->nbclient<0.6.0,>=0.5.0->nbconvert[webpdf]) (6.1)\n",
      "Requirement already satisfied: jsonschema!=2.5.0,>=2.4 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbformat>=4.4->nbconvert[webpdf]) (3.2.0)\n",
      "Requirement already satisfied: ipython-genutils in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from nbformat>=4.4->nbconvert[webpdf]) (0.2.0)\n",
      "Requirement already satisfied: setuptools in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.4->nbconvert[webpdf]) (57.4.0)\n",
      "Requirement already satisfied: six>=1.11.0 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.4->nbconvert[webpdf]) (1.16.0)\n",
      "Requirement already satisfied: pyrsistent>=0.14.0 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.4->nbconvert[webpdf]) (0.18.0)\n",
      "Requirement already satisfied: attrs>=17.4.0 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.4->nbconvert[webpdf]) (21.2.0)\n",
      "Requirement already satisfied: packaging in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from bleach->nbconvert[webpdf]) (21.0)\n",
      "Requirement already satisfied: webencodings in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from bleach->nbconvert[webpdf]) (0.5.1)\n",
      "Requirement already satisfied: pyparsing>=2.0.2 in /Users/eduardo/.local/share/virtualenvs/eduardo-BbA8MaND/lib/python3.9/site-packages (from packaging->bleach->nbconvert[webpdf]) (2.4.7)\n",
      "Building wheels for collected packages: websockets\n",
      "  Building wheel for websockets (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25h  Created wheel for websockets: filename=websockets-8.1-cp39-cp39-macosx_10_9_universal2.whl size=68291 sha256=fcfd7b81c69e2baf0fc87a08c6471a018d22b693ef83692849470d73adc8b7ab\n",
      "  Stored in directory: /Users/eduardo/Library/Caches/pip/wheels/d8/b9/a0/b97b211aeda2ebd6ac2e43fc300d308dbf1f9df520ed390cae\n",
      "Successfully built websockets\n",
      "Installing collected packages: websockets, tqdm, pyee, appdirs, pyppeteer\n",
      "Successfully installed appdirs-1.4.4 pyee-7.0.4 pyppeteer-0.2.2 tqdm-4.62.2 websockets-8.1\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install nbconvert[webpdf]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7b2dfadc",
   "metadata": {},
   "outputs": [],
   "source": [
    "import nbconvert"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c859a41c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "oldHeight": 122,
   "position": {
    "height": "172px",
    "left": "1071px",
    "right": "20px",
    "top": "120px",
    "width": "349px"
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "varInspector_section_display": "block",
   "window_display": true
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}