p1-Copy1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import os\n",
    "from sklearn import datasets\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import itertools as it\n",
    "from sklearn.neighbors import LocalOutlierFactor\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV, KFold\n",
    "from sklearn.discriminant_analysis import (LinearDiscriminantAnalysis, \n",
    "                                            QuadraticDiscriminantAnalysis)\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "from sklearn.utils import shuffle\n",
    "from scipy import stats\n",
    "from statsmodels.stats.multicomp import (pairwise_tukeyhsd,\n",
    "                                         MultiComparison)\n",
    "from sklearn.metrics import roc_curve"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cargamos dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Cargamos dataset Iris\n",
    "ds = pd.read_csv('iris.data', header=None, names=['largo_sepalo',\n",
    "                                                  'ancho_sepalo',\n",
    "                                                  'largo_petalo',\n",
    "                                                  'ancho_petalo',\n",
    "                                                  'clase'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "output = 'clase'\n",
    "y = ds[output].values\n",
    "X = ds.drop(output,axis=1).values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocesado de datos"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Comprobamos que no haya \"missing values\"**  \n",
    "Dado que no los hay, seguimos adelante."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "attr = []\n",
    "# Iteramos los atributos del dataset\n",
    "for i in ds.columns:\n",
    "    # Comprobamos los valores de cada atributo\n",
    "    attr.append(np.all(ds[i].notna()))\n",
    "# Ningún atributo tiene missing values:\n",
    "np.all(np.array(attr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Comprobamos que no haya variables con \"near-zero variance\"**  \n",
    "Vemos que no hay ninguna."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([], dtype='object')"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ds.columns.drop(output)[ds.std() < 0.001]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Comprobamos que no haya variables iguales**  \n",
    "Vemos que no hay ninguna."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "resultados = np.zeros((len(ds.columns), len(ds.columns)))\n",
    "for i in range(len(ds.columns)):\n",
    "    for j in range(len(ds.columns)):\n",
    "        resultados[i, j] = np.mean(ds.iloc[:, i] == ds.iloc[:, j])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "iguales = []\n",
    "for i, j in it.combinations(range(len(ds.columns)), 2):\n",
    "    if round(resultados[i, j], 4) == 1:\n",
    "        iguales.append(ds.columns[i])\n",
    "        print(ds.columns[i], ds.columns[j])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Comprobamos que las clases no estén desbalanceadas**  \n",
    "No lo están."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = np.array([i for i,j in enumerate(set(y)) for k in y if j == k])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Proporción clase 0:  0.3333333333333333\n",
      "Proporción clase 1:  0.3333333333333333\n",
      "Proporción clase 2:  0.3333333333333333\n"
     ]
    }
   ],
   "source": [
    "print('Proporción clase 0: ',np.mean(y==0))\n",
    "print('Proporción clase 1: ',np.mean(y==1))\n",
    "print('Proporción clase 2: ',np.mean(y==2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Desordenamos los datos**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "X,y = shuffle(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Detección de Outliers\n",
    "Visualizamos de distintas formas los posibles outliers  \n",
    "y vemos que se puede considerar que existen pero son pocos y no los tratamos."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = LocalOutlierFactor(n_neighbors=20, contamination=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = clf.fit_predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_scores = clf.negative_outlier_factor_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzsnXd4FFUXxt+bhE5CAoGQQi+i9CK9KIrAB9IRqaIURRRFxYoRIgqoiCBFqRYQEFQUsICC9I5IR3rvJYSQhCR7vj/e3Wyv2VDC/T3PPtnszNy5Oztz7rnnnqJEBBqNRqPJXgTc7g5oNBqNxv9o4a7RaDTZEC3cNRqNJhuihbtGo9FkQ7Rw12g0mmyIFu4ajUaTDdHCXZMlKKVEKVX2dp1TKfWFUurdW3n+OwGlVIRSap9SKlcWnyePUmq/UqpQVp5H4ztauN8jKKWOKqUevd39MKGUaq2U2qSUSlRKXVJKzVZKxXhx/N9Kqb7OtovIcyLyvn96mzFwJCqlrhtfVzPZ3qNKqaN+6p4lbwOYJiIpxvOsUUr1dtKH3Eqp0Uqp40qpJKXUf0qpV5VSymKfNUqpZIvvfV0p9aCIJAH4GsDrWfAdNH5AC3fNLUcp1QnAdwA+AxAOoCKAFABrlFJht7lvQS42VxWR/MZX6C3rlAMc9VMplQdATwCzPTheAfgBQBMALQAEA+gN4HkAY2x2f87ie+cXkc3Gz2cDeFoplcPnL6LJMrRw10Ap1U8pdVApdVkp9YtSKspiW0Wl1DLjtnNKqbeNn9dWSq1XSl1VSp1RSk1QSuX04FwKFB4jROQ7EUkSkbMA+gK4DmCwcb9hSqlZFseVNGrPQUqpDwA0AjDBqElOcHCer5RSIyz+b62U2m7s7zqlVBWLbUeVUm8opXYASHQj4G3PU0gp9atS6oJS6opSapFSKtpm+1fGa3RFKfWDUqoAgEUAiltow0WMmvR4476nlFKfmq6pSdNXSr2tlDoLYKqD7tQDcF5EznjQ9ccANAXQQUT2iEiaiKwD0AvAS0qpUu4aEJFjABIB1PbgfJpbjBbu9zhKqaYARgJ4AkAkgGMA5hq3BQP4E8DvAKIAlAXwl/HQdFAQh4NC5RFQ63PHfQCKA5hv+aGIGEBNspm7BkTkHQCrAbxg1CRfcPMdqwOYAeBZAIUAfAngFxu7dFcArQCEikiaB9/DRAAoaIsDKAEgFcA4i+3fAcgJ4AEARQCME5F4AI8DOG6hDZ8HEAugFoAqAKoDaADgLYu2YgDkN57L0bWuDGC/h/1uBmCdiJy2/FBE1gI4Cwp+T9gLoKqH+2puIVq4a7oDmCEi24x22rcA1FNKlQTQGsBZERkjIskikiAiGwFARLaKyAajxncUFJhNPDhfuPGvI+3yjMV2f9IfwJcislFE0kXka9AMVNdin/EicsJoS3bGNqPmf1UpNR4AROSCiPxknIFcA/AhjNdBKVUMHPQGiMgVEUkVkVUu2u8OYJixzfMA4kAzi4k04/abTvoZCiDB5ZUwEw7HvwFg/ztMsvjem2z2TTCeV3OH4fH0U5NtiQKwzfSPiFxXSl0CEA2gGIBDjg5SSpUH8CmoaeYF76WtHpzvovFvJIAjNtsiLbb7kxIAnlJKvWjxWU7wu5s44UE7NUTkoOUHSqn84NrBYzALuWDj32IALho1dU+IAmdOJo6Bv4OJcyJy08XxVyzO7Y6L4OzAEba/w/Mi8pWTfYMBZGpxWZM1aM1dcxoUfgAApVQ+0HRxChR4pZ0cNxnAPgDlRCQE9NJQTva1ZD+AkwA6W36olAoA0BFms08iOGiYKGrTjjfpTE8A+EBEQi1eeUVkjo/tWTIEQCkAtY3XwdKccQJAuFIqxMFxjs5n9VuA5pdTXvRxB4DybntM/gRQ33J9BQCUUg3Aa73Cw3buB/Cvh/tqbiFauN9b5DAu2pleQQDmgB4P1Yw26A8BbDSaWhYDiFRKvayUyqWUClZK1TG2FQzgGoDrSqkKAAZ40gFhjunXAAxVSnUz9qMogGkAQgCMNe66HUBjpVRx4wLkWzZNnYPzgceWqQCeU0rVUSSfUqqVcU0hswQDuAHgiqLPd6xpg4icAIXoRKVUqFIqh1KqsUX/w236MAdArFIqXClVGMC7AGbBc9YDKGy8npbY/u45APwBYBWAH5VSDxgXqusB+AbABBE57O5kSqni4BrAZnf7am49WrjfW/wKIMniNUxE/gSFyA+grbUMgCcBQEQSwIW3x8FFtgMAHja29RqAbqDNdSqAeZ52QkTmgbbkwQAuAdgDIA+ABiJyybjPMmObO0Bzz2KbZsYB6GT0QBnv5nxbAPQDMAE0XRwE3f78wacAChi/xzoAv9ls72H8+x8o0F809mkXeM2PGm3ZRQAMB7XgXeD33ggudnuEcc3kW9B2b8kUWP/uU42DbDtwYXop+Dt+A+ALAC97eMruAGa6MRVpbhNKF+vQaLIPSqkIAH8DqGYKZMqi8+QBZ1cNRCQr1kk0mUQLd41Go8mGaLOMRqPRZEO0cNdoNJpsiBbuGo1Gkw25bUFM4eHhUrJkydt1eo1Go7kr2bp160URKexuv9sm3EuWLIktW7bcrtNrNBrNXYlS6pj7vbRZRqPRaLIlWrhrNBpNNkQLd41Go8mGaOGu0Wg02RAt3DUajSYbooW7RqPRZEN0sQ7N7UMEWLUKWL0a2LoV2LkTuHYNUAoIDgaqVAFq1gQaNwYaNuTnGo3GI7Rw19x6EhKAGTOASZOAnDmBFi2AJ58ERo4EwsK4z5UrwPbtFPrPPQcYDMCAAcAzzwD589/e/ms0dwG3LStkrVq1RAcx3YP8+SfQty9QuzYwaBDQoIF7jVyE2v24ccA//wDTpwMPP+z6GI0mm6KU2ioitdztpzV3za0hPR0YPBhYuBCYMoXauqcoRdNM48bAkiVAz57AE08AH38MBAZmXZ81mrsYvaCqyXrS0oBu3YA9e4AdO7wT7La0akXb/D//AL16cdDQaDR2eCzclVKBSql/lFK25c6glOqtlLqglNpufPX1bzc1dzUDBgBXrwKLFwOhoZlvLywM+PVX4OxZmnY0Go0d3phlXgKwFyxi7Ih5IvJC5rukyVb8+COwciU17dy5/ddunjw08VSrBixaBDz+uP/a1miyAR5p7kqpGACtwAr1Go1nXLwIvPACMHMmkC+f/9sPDubi6oAB9K7RaDQZeGqW+QzA6wAMLvbpqJTaoZRaoJQq5mgHpVR/pdQWpdSWCxcueNtXzd3GRx8B7dvTIyareOgh4H//A8aMybpzaDR3IW6Fu1KqNYDzIrLVxW6LAJQUkSoAlgH42tFOIjJFRGqJSK3Chd3mmtfczSQlUWMfPDjrz/XKK8C0acDNm1l/Lo3mLsETzb0BgDZKqaMA5gJoqpSaZbmDiFwSkRTjv9MA1PRrLzV3HwsWMLq0bNmsP1eFCkDFisBPP2X9uTSauwS3wl1E3hKRGBEpCeBJAMtFpIflPkqpSIt/24ALr5p7mb/+oknmVtGhA8+p0WgAZMLPXSkVp5RqY/x3kFJqt1LqXwCDAPT2R+c0dzFbtwK13AbR+Y+aNXlOjUYDQKcf0GQFKSn0Z796FciV69acMykJKFiQicdy5Lg159RobgOeph/QEaoa/3PtGl0ffRDsU6dORbFixTB16lTvDsyTh0nIEhO9PqdGkx3Rwl3jf9LTfc75EhcXh5MnTyIuLs77gwMDdToCjcaIFu4a/5MvH3D9OrM5eklsbCxiYmIQGxvr3YEGA3DjBpA3r9fn1GiyIzorpMb/BAcz/8uRI0Dp0l4d2q9fP/Tr18/7cx44AERF0Tyj0Wi0cNf4gfh4YPlyeqscOACkpgIBAcDw4UD//kC9evw/K9m6lR4zGo0GgDbLaDLDsWOsklSyJHO0BwQA7doB3bsDDz5orqJ0333AZ59lbQTp8uVA3bpZ175Gc5ehNXeN94hQmA8dCjz7LPO0R0Za71OnDmugHj0K7N4NxMUBX30FfP01ULWqf/tz9Srwww/AXh07p9GY0MJd4x0iwMsvA3//zVfFio73i4lhKbxZs4Dnn2f+9a+/Bpo1A+bMAR55xH99+uorFgApWtR/bWo0dzlauGu8Iy4OWL8eWLUKKFDA9b5vvAG0aQN06gQUKQL07g2UKsX/ly1jLvbMcvYsC2v/9lvm29JoshHa5q7xnE2bgEmTWBzDnWAHWAS7Z0/mdDfRpAnT8z71VOZt8CK06fftC9Sokbm2NJpshhbuGs8wCdKxY4GICM+Pi4ujzf2zz8yf9exJs82ECZnr0yef0N3SW594jeYeQAt3jWesW8fQ/ief9O64PHloMhk7Fvj8c36mFPDeexTuBlf1X1zw6aecRSxZcuvy12g0dxFauGs8Y+ZM+qz74q9evDjrqH7+OdCnD/3iH3yQpp1Vq7xr68oVmnSmTuWxMTHe90ejuQfQwv1e4swZ4JdfgG++ocfKX395bvfeuJEl7XylZEn6vefMCVSuDMyfDzRuzHY9ITUVmDuXxwYHA5s3A8UcVnPUaDTQ3jLZHxFquBMn0kOlXj2mxk1NBY4fpx96377Uyp0Jy5s3GXlaqVLm+hIcDEyezICj996jLT46GmjalEI7d27r/ZOTgR076EY5dSpQpgxdKzMzyGg09whac8/OXL5Mwfn88/RSOXaMgnLWLGDePLo0Ll/OFL3VqgGjRztO9nXjBu3aHti2PUrZ27QpsHo10xNcuwb068cBp0oVavONGlHYFyzIIKnLl4E//uAgpQW7RuMRulhHduXiRQrJ1q0ptN3Zyk+dAlq1YnDRJ59w0dPE9ev0U09MtP7cAcWKFcPJkycRExODEydOuD7njz8ysOnnn1lsY98+CnulqOXff7+9Nq/R3ON4WqxDm2WyI6mpDB5q144BPp4QHc2I04YNgRIlgEGDzNvy5QNCQoATJ7g46oLY2FjExcV5lrJ33z6gXDm+z5MHqF7dfp/kZGDtWmDLFppo4uOZtz0ykonC6talucjNoKPR3GtozT07Mm8e3QxXrvTeu+XgQQrMY8co1E20agU88wzQsaP/+tm+PfDEE0DXrvbbTp7kd5gxAyhblrlqqlWjqcZgYP+2buV3LFSIpqdevXSJPU22R2vu9zKTJgEvveSb22LZskCDBvSm6dvX/HmbNsDs2f4T7pcuAStWcKHUEhF+9s47QI8e1NpN2r0jDAZg6VKakiZMoMumP9IaaDR3OXpBNbuxaxe177ZtfW/j+efpXWM5q+venWabw4cz3UUAzCrZti0QHm7+LDmZg8fUqRT8Y8e6FuwAB7AWLegJ9NJLTEw2c6Z/+qjR3MVozT27sXQpzR2ZMU80a8ZI1IsXgcKF+Vn+/MDrr9N7ZenSzNm4Dx1ifpl168yf3bwJdOjA86xdS394b1CKicnq1WP/DQYGTGk09yhac89uXL7sNveLW3fFgABq1FeuWH/+2mtc0PzoI9/7l5jIWcDbbwPly5s/HzqUf2fP9l6wW3LffcCff7J9vaajuYfRwj074maRPC4uDidPnkRcXJx3bQQF0X3xyy+d+8S74vJl4H//Ax54gDnhTWzYwKjZr77yz4Jo+fI06fTuDaSkZL49jeYuRAv37EahQsxx7oLY2FjExMQ4d1dMTwcuXKBnii0xMQwmmj+ftu5jx9z3SYRpDypXZk6ZadOsF3sHD6aZpkgR9215SteuTHngKphKo8nOiMhtedWsWVM0WcC+fSIRESLJyb63sWiRSO3arve5eVNkxAiRsDCRzp1Ffv5Z5PRp8/a0NJFdu0QmThSpWlXkvvtEVqywb2frVpHixbm/v1m5UuT++0UMBv+3rdHcJgBsEQ9krNbcsxv33cegnh9/9L2NSZPoMeOKHDnornj0KFMGTJjA84aG0uYfHMwgqrVrgY8/Zp1VR6kDZsxg+oHAQN/764xGjbjQun69/9vWaO5wPPaWUUoFAtgC4JSItLbZlgvANwBqArgEoIuIHPVjPzXe8PzzXPR84gnvhebevcy4+MMPnu0fEsJKSy+8QPPLlSv0fAkOtg6Ccsb69XS7zAqUYh6b9euB+vWz5hwazR2KN5r7SwCclZfvA+CKiJQFMBbA6Mx2TJMJ2rShYB082LtFzwsXeOzIkUwH4C1K0U5ftKhngj0lhYNJ1aren8tTatZkJKtGc4/hkXBXSsUAaAVgmpNd2gL42vh+AYBHlNLJPm4bQUHUvNesAQYMANLS3B9z5AjNGN27W0emmoiPBwYOBGrVYkKvWrWAF18EEhJ87+eFC0BYmFcDiUdZJy0pVYqpDDSaewxPNffPALwOwFlNtGgAJwBARNIAxAMoZLuTUqq/UmqLUmrLhQsXfOiuxmNCQxlReuoUUwqMGgWcP2+/3+bNwNNPs8D0iy8Cw4ZZbz96lIFBYWEslhERwX0jIuiTHhrKdAXuMkA6Ii2NA5EXeOTGaUlgoGeDm+beIjmZ9/633zJaesYMppW+eNHzNtLS6Jl24gTTadymPF3OcPtkKaVaAzgvIluVUg9l5mQiMgXAFICJwzLTlsYDQkKARYtolpg0if7fpjzpN2+yWMeNG9TuP/rIHI1qYv16LoKWKME88C1a2J9j8WKG/Zcrx5lCLbf5jMwEB3NGIOJxxKtXWScBth8c7HmfNNmXGzeYM2naNODff3nPVqwI5M1LQX3sGLBtGxWZJ59kNHapUubjRcwDwqZNTPWRPz+dCxITqajUqEEHg6efBqKibt93ZX9du9MAGAngJICjAM4CuAFgls0+fwCoZ3wfBOAijBknnb20K+Rt4PJluiP++KPI4sUiGzc6d0E8dEgkZ06Rli3dt5ueLvLIIyK5cokcO+ZdnyIjRY4e9e4Yb/jgA5FXX8269jV3PunpIp99JlKokEjr1iJLlojcuOF83507RQYPFilYUOTJJ0XOnaOrb82aIqVK8Z5atUrk2jXrY0+dEvnlF5HnnqOLcKdOIgcO+P3rwENXSK980wE8BGCxg88HAvjC+P5JAN+7a0sL9zucWrVEypf3fP/0dJHSpUXq1/fuPK1bi8yf790x3tCunch332Vd+5o7m0OHRBo25GvfPu+OvX5d5MUXRXLnFilalPEf6emeHRsfL/LRRxxQxo71/DgP8FS4++znrpSKU0q1Mf47HUAhpdRBAK8AeNPXdjV3AJcv05QzebLnxwQEAOPHM5WAN4usjz/Osn9ZwaVLXHdo1ixr2tfc2fz7L4vPtGvHvP/33efd8WfP0uzYti1NL//953ka7ZAQYMgQmjbnzwe6dLn1qTA8GQGy4qU19zuYfv1EwsN9OzY0VOSFFzzfPyFBpEABkT59GMlarBhfVavSnJKZae1HH4n06uX78Zq7l//+o8lv7lzfjj92jJHTkyfz/+PHRcqUEZk0yfu2kpNF2rYV6dBBJDXVt/5YAB2hqvGZDRvoFukL9et7HhG6ciXT/KamsmD21Kn8u2oV3wcGsr0WLfi5N5w5wwIer7zi/XfQ3N2kpQHdujEzaJcu3h+fns4AwBdeAJ57jp8VK8ZU1++9x3KP3pArF6ujXb6cuYyqXqKFe3bh1Cm6MZYrxwCifPnoHTNihNtEYnbcuEGPAQf07NkTQUFB6Nmzp+NjQ0NZUNsdU6fSI6FXL+D0aQryQ4fomVOyJBOMjR5Nj56uXYHOnemu5gkifCj79cvaACnNncnHH/M+HDjQt+PHjKEHzauvWn9eujRdinv3pkLiDblysRj82LH0srkVeKLeZ8VLm2X8RGoqzSBhYSLPPy+ybRtNHQkJIlu2iDz7LE0lr7zieXKuypVF2rd3uCkwMFAASGBgoONjW7WiV4ErvvpKpEQJa5PLli00Ba1e7fiYffs4TZ41y3XbBoPIkCFMfJaZ5Gmau5OLF3m/++qBde4cn6XDhx1vNxhEmjYVmT7dt/a/+ILHZwJkhbeMP19auPuBtDTa8Zo3F7lyxfl+ly7RVfHJJz1btX/6aZEiRRxu6tGjhwQGBkqPHj0cH1uwIF3BnLFjB9veu9d+29KlFPALFjg+dtcukcKFRXbvdrw9MVGkf3+R6tX5kGvuPT75RKRnT9+P//BDkWeecb3Pr7+K1KjhW7bRlBR63uzZ41v/RAv3e4PYWArtlBT+v3q1SJ06Innz8lW/vsi6ddyWlER3sA8+cN/u+fMiStGX1xt+/53HXb3qfJ++fV33YeNGkXLlOBCdOWO//b33RAYMsP/877/Nx7k6vyZ7U66c+Z73FoOBM8qtW13vl55Of/fNm307z9ChIoMG+XasaOGe/UlMpA/tkSP8f/RokYAAkWrVRCZMEBk/nuaVgACRceO4z7591Jo9MVdUry5SsaJ3fSpfnoOLM65c4ZT57FnX7SQmirz2Gvft2lXkt984+xAROXmS0+b4eJGDB0WmTqVPfqlSIj/84F1/NdmL06c5c/Q1f//BgyIxMZ7tO3CgyJgxvp1n82aRSpV8O1a0cM/+TJ/OACARumkpxSmpLe+/TwF/7hz/f+QRkdmz3be/b59Ijhw0+3hCq1aMaHVmqxThINO1q2ftiTCi9rPPRBo3FgkOps29alW6TubJQ5fJTp0YcZgVxT40dxeLFok0a+b78fPm0WXRE2bMEOnWzbfzJCfz/k1M9OlwT4W79pa5W1m4EDB5rLz+OtPs2q7uAyw8XbAg8NZb/L9nTx7rDlOh6UWLWIRjzRrH+/39N1ChArBsGd9b5uKwZds24NFH3Z/bRFgY89asXAlcvQosXw7MnMkEZx060JNm/nzWZc2KYh+au4s9e5g7yVd27waqVPFs38qVub8v5MrFZH779/t2vIdo4X63cukSEBnJ9//8w2RFRurWrQulFOrWrcsPGjQAtmzh++hoHusJjRsDO3cCOXPyfUQE0KkT0KcPhWuRIiyGkT8/H6x69Vy3d/UqXdRc4DSlb0AAUKYMUL06Hyxd+Fpjy/XrPiWJM91zO9avZ2SpJ4SEANeve5+C2ub4rEQL97uVXLmY2RGgT65FqtKNGzda/cXFi+ac6SkpPNZT7ruPGvf580CrVgzB/vtv4OBBhmVfvMiBo0wZ7/rsBI9S+nr7HTT3BoGBDEDyEtM9t3b9es+PN6ar9joFtc3xWYkW7ncr5cqZtfE+fRjVeeMGAKBOnTrmvwkJjDg11UTdvJlTQm8JD2eujI4dKeQ7deL/BQt63kbhwm7zvsfGxiImJsZ1St8TJ9gfjcaSiAgGxHmJ6Z6r3rKl54VdTp8GihTx7H51xMmTnPlmJZ4Y5rPipRdUM8nmzfQQSUvjq1Ah/n/gAF0i16wR2b+fi44REXTfunlTJCqKvua2pKYyD8cjj9AdrGpVkZEjRS5coGtky5Zs59VXmeVu8GB63jz+uNmTxR1//UUPHF+9GUR4bPnyzoOdNPcuGzfyvvWVFStE6tXzbN9Ro0Reftm385w9S08wH58DeLigmrXzguzI2bO0cV+4wGlgmTIMcfel5mhmqFWL2uuPPzI0f8cOFh4oV44Z7ACGSBcsyDqlAQHMvlimjP2iU3IyM+ddu8ZcLDVqMDfL9On8bvnzc/vChbS/mxg5kgu1jz7KmUP+/K77/PDDnPauXm21RuAVy5fTJNOggW/Ha7IvVarQbHj9uvt70RE1anCNKSnJ/fO8cSPXnXxhwwbW9s3qSqSejABZ8bqrNPf4eLrkVahAH+tHHxXp0YMBMzVq0K3p8cfpj50ZrdRb1q5lxObGjXR5rFGD7lzPPsso0fnzRapUEfn4Y2ry4eEimzbZtzN4MF0KHWWsa99eJH9+59nsDAaRjh3pZ+8J48czx7ov18lgoMulKVOf5s4nKYkZFnfv5t+kpKw9X9u2Il9+6fvx//ufyMyZrvc5f57uuL5GQbdpw/gMH4H2c/cTixeLREeLdOlCU4AjoXT9usi0aSIPPMCb4+TJW9e/RYsYuJEnDyvI2LJtG6NVCxUS+eMP++3XrnHAOnHCfltaGoM6KlZkhRlnbNwoUrKkZ6kNEhI44Iwa5X5fW0wDmI/+wZpbyPbtVDIKFODzU6EC/xYowBQR27dnzXn/+IP3l69K1uLFDIpzdfyoUSK9e/vW/rFjfF6vX/fteNHC3T988AHtz8uXe7Z/SorIsGHMHfHPP1naNSvGj6c9PDRUpHdv2dS2rYwOCZH99erxs8KFnWsKK1aINGhg9dGUKVMkJiZGZo8aRRv96NEZ9kXTtilTppgPMBg4K3AXeWri5EmuDwwf7tlDaDCIvPsu82mfPu3ZOTS3hyNHGHQWHS0SF2f/e505w0E6JkakUSNzhLW/SE+ncHeXYM4ZaWk8/ptvHG8/d47Pmi/Pt8HAmfC77/rWNyNauGeWCRNE7rvPN2Eyfz4FvKtoTX+ycCEXPM+cERk/Xj4PDpYPAHknNJRTyKZNaTJyxF9/8WG0ICYmRgBInaJF+RCOGZORC8O0LcY2TLtIEe+u1Zkz1JBq1eI02FFNy8RERuLWqMG0BqYoW82dyY4dVAY++cR9UYrUVJFPP+X+//7r335s3sz70VFuIk/Yto0K0alT1p+bTJBvvOFbu999x9l9JrOVauGeGfbsoRnj0CHf2/joI5GHHvJr7USnXLjA6e6FCyJio12fOUPt3VkyrYsXrY61PH7apEm8yevUEZkzx75tE7t2seqNt1Vm0tI4Df7f/6j59+hB+//LL4t0724uaPzrr7fmOmp858QJKgKepLawZM4cHnf8uH/7M3SoSJMmzgthu+P990UefNC6CPa4cRTOvqwbbNrEZ8nXZGMWaOGeGZo1E/n888y1kZbGnOLe3uy+8tJL7Pfly+bPLl7kAONO03jmGS7AOjKRdOpEm70p86Qj+vZlhsrMcOgQtfRPPuFr+vRbN/PRZJ4nnmDGTl8YPlykc2e/dkfS0ujw0KyZb/Ztg4HZR+vU4ez388/pVuxLnviVKzmTcLVu5QVauPvK3r20qfmj0MPChZ77zWYWU9GO0FDa9dq14/vBg90n1bpyhaaP9u2ZLjUlhTfx0KHUnosWpV3fVvgbDDTZlC5tpflr7jFOn3Y9O3RHfDyPtzWDZJbUVNbmvf9+kQ0bvD/eYGBcR548nJl6uz6QlCTy+uuUJ0uXen9+J3gq3HWEqi0/yE+zAAAgAElEQVTz5gE9evgnvL1VK+DwYeDo0cy35Y6gIODzz4F9+1j/sUsX+vx++qn7pFqhoUzO1bAh8PTTLNFXqxZzwWzYAKxdC0ybRj/isWOBBQvYbqVK9J1fvlxHjN7LTJvG+61AAd+ODwlhycVp0/zbr6AgtjlsGFNlvPgi02Z4wvXrwJQpwJw5QOvW9Jvv1w/45Rf3KQquXgXGjePzcegQY1CaNcv01/EWxYHg1lOrVi3ZYgqfv5No3Rp45hnfAxRsadOGdUI7dfJPe+fPs/DvpUvAa6/dumAeEQrxuXNZ6Dc8nLVNmzTJ+mAMzZ3NffcBs2dTIfCVbdso4P/7z3/9suT8edZW/eorBis99hgDiSpWpDKTmgocOwZs3coC7wsW8N4eMoRF2pOTmYF00iTuV78+jy9VikGDiYnMErl1K18tWzLlR8OGfv8qSqmtIuL2Ymvhbku5ckxzW6GCy92mTp2KuLg4xMbGol+/fs53fOst3jxDh/qnfy1bYndKCmZu3473U1OR59gx7/K7aDT+JjiYuVJcaO5un5dr14CoqCzPlIikJEZar1tHIbxvH3My5cjBjKk1a3KQ6tQJKFbMcRsHDjBH09at/N6pqYxorVCBx9epAxQqlGVfQQt3XylRgiaKkiVd7lasWDGcPHkSMTExOOEqGdawYYDBAHibNc4ZJUvioeRkrDx3DoeDglBq27bM5bDOCkS0Nn8vERREAWmZmsIGt89LaiqQOzezJep7xyWeCndtc7elQAGP8p17nA3u0iW3Ocy9ondvLASwKlcu5IuJcTvDuCWkp3O207IlZymBgczQ99JL1Iw02ZsCBYD4eJe7uH1e4uPZjhbsfkMLd1uqVWNiMDf069cPJ06ccG2SAdhWtWp+6hyA995D6PjxaDRmDIps325OEna7uHCBdsX336cN/vRpamHr13O63qQJEBtLbV6TPalcmYnjXOD2eVm16s6bgd7laOFuS8OGwB9/+KetS5eAXbtoh/MHP/4IPPggMGAAMGoUcP/9NPe4KYCRZVy/DjRvTgG+cSMXjgsUoOZeujQwYgSz7C1eTOGvyZ489xwXGjPDpElsR+M33Ap3pVRupdQmpdS/SqndSqnhDvbprZS6oJTabnz1zZru3gKefJK1Q31I+m/HjBl0wfLVRcySCRNYI3X4cK78nzjBuqWbNtGzJy2NFYp696YnS+HCQN++1KJNXLhAD5vKlTmIzZyZOY160iR6C4wc6Xw6XaQI8NtvdA07ftz3c2nuXDp0oKfI3r2+Hb9vH5Ugf3moaYg7R3gACkB+4/scADYCqGuzT28AEzxxrDe97tggJhGRIUN8r2xu4swZRqX5I/vd+fMM8jh8mIEiQ4eKDBzInBw3bzJQatYskbJlmZ73tddEXnmFkaUPPMA2EhKYmW/AAJEtW0SWLGHg0ptv+tan9HQm/9q40bP9Bw0Seecd386lufMZMULk4YddRzI7IiWFBWLefz9r+pUNQVZEqALIC2AbgDo2n2cv4Z6YKFKuHBP9+EJqKvOl+EuYffqpOcVo48YiPXsyY2Xhwkwh+sMPIpUriwQEWCfXOnlSRCkmB5s4kRGolpgGjfPnve/T1q1MrOZpatWtWzm4aLInaWmMiu7UyfPcK8nJTDvQtq37KGpNBp4Kd49s7kqpQKXUdgDnASwTkY0OduuolNqhlFqglHLiIHqXkDcvAxZefpnBDN6QkgL07Ekzibd1FZ1x+DAXZW/cADZsQC+DAUGxsdicOzcjSKtVYxRsdDRQpAh69uyJoKAg9HzzTaBoUWDJEgaJtGgBwFztferChYyi27PH+z5dvky/ZBtzjNNq8JGRwJUrPl4AzR1PYCCjOZUCmjZlwJszk58IsGIF9zMYeJy7KGqN93gyApheAEIBrABQyebzQgByGd8/C2C5k+P7A9gCYEvx4sVvxSCXObZtEyleXKRfP+ZfccemTSKVKlF7cZSNLjWVWvbTTzN1aJ8+Ij//7F5rGT6cmRINBpGyZeU1paQlIGcAFuj46y/mnc+RQyQlRQIDAwWA5FFKJCiIJpsPPmDxBDGn7S0THc1sjL7k1F6/XqR6dbuPnaYE3ruXOWg02Zv0dJEvvmCBlwoVWMHs7785c/v7b2ZWrFCB2ydP1tk+fQBZlTgMQCyA11xsDwQQ766dO9osY8nVq6wcExZGe/WvvzKD4fbtLB22e7fIV1/R3li0KAWpI1PFvHksYNCwIU0k8+Yx01zduhxAfvrJeR8OHKAQjo8X2b9ftkVHy2pAxjVqxHN16MCi1QUKiJQtKyMefVQeU0rO5cvHqi8GA9cAIiJEvvxSpk2cKLUiI+VgrVrM5ucpBgPTIf/9N+394eHsmwUOUwKLsHCDcXC5bVy6xKLhnhb0tiQpiaUKly93nCc8IYFJ19at4/t7HYNBZNUqkV69eM9Xrcq/vXoxS+KtLEeZzfCbcAdQGECo8X0eAKsBtLbZJ9LifXsAG9y1e9cIdxMnTzLDW1QU7dicXFIzrl6d9nlni0nTpzNntbPMdKtXs11X6YEHDmTlGsuUo9evM51vxYrMO330KKvIKMVX9erWmfa2b2f916AgkeBgLnJ6WrJu8WIuwBYqxIpIpmyRbdu6PzYpid8/q0qruWPLFvYzJISLziEhtA9v2+b+2JQULmAXLszCIk2acKDv2JEL3PHxzMYZFmYuPhIWJvLii9a5wDUaP+FP4V4FwD8AdgDYBSDW+HkcgDbG9yMB7Abwr9FsU8Fdu3edcD97loJh8GBqbqmpNKesXcsHeuBAx9rIwYMUhPv3u25/505q2Y5qmYrwXLGxFExFi/KVLx+FluUi6qZNFPhvvsmpsCUGAzXLoUOZM93TSjXffEMtvWBBDjADB4rUr08hFhBA4eaMlBTOLLp39+xc/mb5cgrmyZPNeb0TEjh7Cg/nLMQZqaksfN6qlfUMJT6eZq7ISC5kP/209e92/DgXwB98UGvxGr+TZWYZf73uOuH+1FN0kXREfDwFv6Naq6+9Zn3czp00xRQqxNqle/eatw0c6LroxZgxIrlzi+TMScGeOzc1SVMFpPffp+nn3XcpwKOiWBFKhIJ94EC6L771Fgt0FCok8uefrr/3lSscUMLC7Pf97TduU4rXx7Ku5M2bIt9/TwHXrl3WV713RGoqZwzLljne/vvvNIk5W/P49lu6md686Xh706a83o4GdYOBniDDh/vWd43GCVq4+5NLl+xK0dkxYYJ9NRmDgQL04EFzO1FRLFZ99iy1x+LFOTiIUPDbLkSauHBBJFcukRYtuFhrMPD4PHloEvrvP2qolkWqT52itn30KAthly1rbSpYupQLsa4WdMePFylf3nk1p8GDOWC0aUNBV7IkzUQFC9Jtc+7c2+fm9uOPdsW/7ahTh4vajqhfnwVXHJGaSs09JIQmO0ds385rot38NH7EU+Gu0w94wt69TNBlUZDCzuXv0Uftc9LcuME8z2XK8P+NG4H770fPlSsRFB2NnuvXAzExdFMEgAceYGSso2IAK1Yw+97o0ejZvz+CcuRAz3XrmGp07lzg99+Bdu2AiAhz35YsYX76P/6gO2SvXkBwsHn70aMsSrJrl/Pvvn073TvbtgUAs5tlz57cbvwc0dF0x1y2DPjuO7a5ciWLONwuN7dt2/i7GHHoptmsmfNcQtu3013P0fEXLvB3qlOHEbixsdjfsCHm5suHba1bMy951aq8dhcvZtU31GicooW7J+TMyTzQFsTFxeHkyZOIM6XyNeWEtiQoiP7uYvT3zZsXuHgRc7/7Dunp6Zj33XfMP5MnD7enp9NPOMDBz5IvH/9euoQ5c+YgPT0dP8yZwyICISHcbvQjt+rb5cusIpM/v932D4YPZx7t4GDn3z1HDgrnhAQAyDj3nDlzuD0hgf3NkYPft2xZVmyKjPTs2mYlOXJY/W52vxng+Hfz5PgcOXjsxo0s6JCUhOm7d+OvGzewaeVKoFEjDiyu2tdoshJP1PuseN1VZpmUFLoR7tmT8ZGdy98bb9BEYUuFCnQJE6HttmFDWVeihPRSSjYVK0bvFdO0/fff6ZHirA+hoSIxMTKiWTOpGBAgm2NiaJZZu5aFsQsVElmxIqNvS15+mYuG166Z3Sm3bpUpU6ZIseho2dymDRdIXfHjjzS1GE1OPXr0kMDAQOnRowdNQ23b0rS0ZIk3V/TWsGkTTUbG62v3m6Wmsuix7cKzic6drQqlWx3/7bcigYE0lV28aL89OZkpIIKCfI909pTNm7lW8+qrjGa2NM1psh3QNnc/Excn8thjjgtnm3y+//vPftu4cazCbuL6dVaJ79qVbVouNLZpQ3u8M9ato+0/Xz7mjcmVi14bJv78k/lsypWju2LRovQpNjF/Pu3yNWtSINet69w7x0RqKu3ykZF0Bb18mZ9fukThFR1NW/6daldu0IB5TxwxbBjXBZyxejXXQGyv0dKlvM6VKvF6mtZMLLl6lW6pH37Ifd0tXPvC/v1cMyhRggPRk0/yHgoNZUyBP4q8a+44tHD3N6mpjDytUYOa2OnTfLhiY/mAf/+94+OuXqUAnD/fdfvffsuH1OSu54zkZLY1ZYq1hnbjBj1gChRgsEjDhnxv+5CbgnF27fI8kOTffzlzKV+eicnKlOHf8uU5gOze7Vk7t4OTJ5kDp107CuVDh0T++INC8P77reMAHDFmDLX7SZM4sCUnczZQpQq9ZQYMYDtffcUkahs38n2FCiIvvcRrvHChSLVq/g3cOXSIA27v3hw8GjWiwlCqFO/RZs1EWre+cwddjc9o4Z4VpKeLLFhAU0pEBLXf5593L9y2baMQjIvLmMJncO4cXRcjI+kt44rERJG+fSlY8+enADe5UnbuzMHn6lXz/leuUIj16uX9d7Xl1CmRt9/mQJUjBzXad9/13Ff+dmLya69dmwNonToU1p76oK9cSV/9vHnNppjJk2kqM3ktBQebg8eCg7ndJMzT05l6Yf16/32njh1FevTg72DpgpqeznNHR1PIz53rv3Nq7gi0cL/TOHiQQrZAAdqp+/RhgEyBAgyCsYw8dUa/fkwXcOkS7fcTJ1JYbdvGwcHRNPzGDc4sHJmMNN7z9NMMADORmkptedo0CvP0dM6qSpc2xx+IiIwezd/PH5w8ybiDihU5C3HEM8/QTNOkie/nOXGCA/hDDzF9QK1azHb61VeOcydpbgmeCvd701tm/366r/3yC7MqOnI99DdlygBffw0cPMiCILVrAz16MOPjjBkszO2KlBS6GE6ciKk//IBipUtjao4cQPHiwNixdDnMlQuAjctenjys5L5oUdZ/x3uBY8fo4mji779ZGKVPH/Ts1QtBOXOi56pVQFgYsHq1eb+qVekq6g82bGB1r0uXMlw97VxUn3qK7phr1nhfkOWff1g4o0oVelu9/Tbv0YkT2e7cubzvhgwBrl513IYIr02XLkD58swgWr488MQTdOv1tk8a7/FkBMiK1y3X3FNSmLulQQMGEjVvzrDyypVpXvnwQ9/ymt8qrl9nRGpysnXmxbZtOUUfOjRjV7vMjEOG8PtpMk+jRgwIM/HTT7yXRDKycQYGBtLm/csv5v3+/JMasD+YM4fBbGXKZHxkdW4RegBVqsT0EN7Y3RcupHPAhAmuc+McPMjZQcWKTLdgybJlXIe4/356G+3dy9nG3r1s94EHuG3pUi++tMYEtFnGgsuX6RXRqBFd+yynyyJMLPXMMzRtbN7svr1//uGi5s8/39oBoWlTkTFjMlzu5sfG0jPi+++tFuysXPIMBj5IrnKoaDynY0eRr782/3/1Kk0kO3dmuIm+2aoVfxdLL5oZM+wjmH1l+3YqJBbuuVYuqiL0BGrXjovenmLyAvLkGRDhvTV6NBesTWtJc+awjcWLnS8gGwzMrhoR4TpZnsYhWribSEyky9+LL7rPHf3TT7wxd+1yvH3RItodS5bkAlvLlnyIu3c3u8vt309NuX172tWXLfNfzuoDB6it1anDWUdoKFMHp6dTgxo3zv6Yjz6i66NOseof5s5lemdL5syhgH/6aXqvhIXZL2Q2buzeY8obatdmKciHH7bP7Pnvv1xnadmSC/5PPME+5cvHXD/TptFrKi2N90+TJpwVAlQSvE0Z8corHLj+/pvPjzvHABO7dnF/y5mQxi1auJt47z1qW54Kt+nT+QDYMmUKPRMWLbK+8S9dokkkJobeC/ny0VsiRw56VYSFcTHKX4ElqalcRJs/3zov+aFD9G9v2pTeEhMn8qGtUIGl+LI7v/3GAa94cQ5m48ZlzaJfSopj98+TJ2mC+Pxz+1wzCxZQeBYtysXX994zxwv4ytKlbO/xx/mdhw/nvdunD+850yBTrBj7dPYszSxLltDby6QgPPgg76WPP6bCMn8+P2vf3nnCNFsSEniumjWduwQ7Y8ECKl8aj9HCXYQ3Z2Skc03cEWlpfAA3bTJ/9t9/DgtTWPHII4xGbNbM7Jp29ixT7wYH07bvD5/jffs4M+jfnwLNkpQUapHPPMOH/PvvPX9A72ZGjOCM5uuvmWN9+XJ6dTRs6Hm+em8YM4aCzBNXyu++o3tkjhy0zdeowUE/Oto84B88SMWgVCkOyIsWedaPuXMpVB9/nApM69b0kKlXj+tKZco4NhsaDMz1HxFhduesUMEc8JaSQsHvLAuqI7p0YRI1W5OnO9LSODg5ixLW2KGFuwi1EF9cwUaP5hTbxODBzrMiitC2mjcvH+AtW+y3x8YyS+Lixd73xZL16znIvP02Ky+VLes8+vJeYe9emiBs/e3T02lzHjXK/+c0GDi41q5tv5hoyf79XNAMC7PW1NeupSb/xBOMc4iKYrrm/fu5JlSsGMsxesK5cyIjRzKDZdWqHNS+/57C+9dfHR+TkECTXsGCHFjWr+e6jOXs9vhx9tvTWIDOnanE+MKHH/rPTfQeQAt3EWrN77/vcJPTcnAi9BuvWtX8f8mSdtNwq+O/+462w9atuZAlNgtcFy9SW+vSJXPfp0ULkZkzM849a/RoPqSWgUv3Gm++aTXwWv0uGzZwsS8rMBg4sJqqMv3+O2cNhw5xRtWhA010OXNmLGZb3RO9e/OeiIvjLMui77++9BJnBr5iil61WOuxui4//MBUGi+8wIFhzhw5WKuW/fPQvHnGIOPyeRGhs0JAgFPN3eXxK1dylqXxCC3cRbiIOnasw01OCzmLUIMqW9b8f8GCdtNbq+MnTKD2OGgQF5fEgWtagQK0h2eGUqVE9u+3Pnf58t6ZnbIbvXtzgdCI1bW5fJnXPSuJj+f6Rr16DCgrUYLvJ03iDDBnzoy1Eat7YvJkau/durGgtEXfK0VGZq7fW7dyYdQCq+syfTqLq3zwAQfHadPk+zx5JMj2eejdm/vaHu+IGjWYxM5Rnh13x2/Z4rDYusYxngr37B3EVKAAgzAcEBsbi5iYGMTGxtpvvHyZaXRNREQAhw6Z/9+/H1Pat0eX8HCMePVVBnSIMH94RAQAoGvXrggMDETXrl2BI0eA1FSgVKnMfZ/69YG5czP6Pv6ppxhEUrZs5tq9mylfnml3jVj9rhs2cHtWEhICPP88sG4dg5SOHuX7AQMYuJQzJ4N5YHNP/PILUyQ3bMj3Ihl9n/DYYwxyc4fBACxdCrRvDxQsyIC14sWBKVPYj+TkjF2trktUFAP5tm8Hdu4E3noLnZKSkARgT1IS8MEHwLlz3Ccqyv54Z9chOZmppR3g8virV62fN41/8GQEyIrXLdHcly9nwIS3boCvvcaXiQ8+4NR50SLrIKjGjWkWefppalpK0Q3NEoOBrpIhIUwqlRmOHKFm2KKFSM+eNAl4652Q3ThzhtfBdkEuIYFuqzNn3pZuiQgXCyMjmYrZtBZz8yarW+XOzVleYiK13g4duBj+7rtcV3FWTN3ExYs0ZVSsSHt3vXo0JT7yCNvKkYMBbo7u/dRU3rdKcebw6qs0IQ0aRNfZPn3MNnlPF0h79OD+vvDKK67XtHzFYGBOpF27GA/gZFZxt4F72iwTH0+7+YYNzPHhTbrVGzf4cJlK44lw0SpfPnoX/PCD9Q1/7pzIO+9w+h0UROE7fTpvpt9+o/dMWBhdy/zha56YSCHwxRfu0/XeK/z0EwXooEF0rfv4Y/7u/fvffv/+f/+loMyTh+sy+fPzXmrUyJzuOSGBedg7dhR5+WX3xdSvX6dNvlMn3lvVqtFkFxFBYV+2LD1lgoJ4DWz56ScmQMuXj+6UjRrx+J49+T4iggpMSIi9R5YzHn+cTgXusprakpjI3+7IEe+Oc0ZKCr2IWrWiqbRwYS4Wly/P71uuHAcvS2+4u4x7T7gfOcLCz/fdx5usShVqbuHhXOhp1Iiucu4KNb/xBjMpWrJyJW+SggXpT2zyzDAY6PnQujUfqCJF+EAULsyHrnBhPtR9+5qTeq1eTY2qQgVqWP7MFHgvc+wY4w3at6fnxapVt1+wm7h6lWs/zZpRIP/6a+YC20aOpK96WBjdCLt2ZRrno0e5uF+5MgV+iRLUzjt25DlXrKCDQWAgfcvz5OH9WqkSlZOAAL4PCeEzNGEChb67vp45wwGsRQvrpGqeMHYsBXFmMRhEZs3iTOnhh3kdTpywvgdSUxlgNXw45UKRIhwMy5Xj3/btOUu3VOzuQO4d4Z6YSG2nYEH+3bbNWrM2GOhyVqUKH4ioKMdFjw0G/ujlytkXwm7fXuTLL+lj3r8/b+SSJanhlC1Lv+cbNzg1nj+fUX+jRnHRzNJFzxRyPXUqp4pffMEb7K+/zPvs3MnPcud2XUhCc/dw6RIFe+7cfGXGfTUtzWyaq1bNsTkjIYECvmxZCs5GjWiuadSIUav330+z4vjxDL6rXZszgVy5KMxjYhj41Lw5P3dVZSs1lVr74MGMB4mIcF5U3Jaff+b+7mYq7oiPp9trpUquNfLERF6vggWpwHXrxmf5jTeY0mHOHDphmKJ7Tem07zDuDeF+6BCnW9262edJtyQ1lWHYRYty6lmiBItY3LzJaeTUqVytf/BBe3/pK1eoyVj6+yYm8twnTlhrNTNnUht3RrVq9g/K/Pn0UTYRFER/4cce48+TmZStmjuDJ54Qee45CubTpzlr89SP3ZZ166hl5s/PdR5nbrBz5nC9qWlTCmsTvXtTew8Pp5b9xBPmbfv2UdjWqUPTVng4XTWfesrxOVJSuJ706KPmYLlNm/icvf8+n48ZM6h09enDv9On8/MPPuB+mV2Hio/nc9u/v+vKU//8Q8XtySetYxOOHuWg8NZb5s+SkhjhXKiQVZnFO4XsL9yPHeOUdOJEz49ZuZLTUaX41U2vEiXYjqPp5/79Vtn3TDj02129mtq7I5KTuciVnm597M2b7I/BwIEFkEa1agkA2RASwqmz5u4mOlrk6NGM331Dhw6O6+16ws8/0+wQGUkhbKROnToCQOqYPjt2jDPAqlWpkZto1YpBQ/XrUykyJkEzHd+1cmUKtdKleeyHH9qbTVJTOTjVrk2N2TYKePVqDmBKMSCrWzeRgQP5t1gxfn7ffebawr5iMFDDHjDAtQlu82ZeC2e1bC9e5EBo64N/6BBnOcbYlTuF7C3c09LotTJypGf7p6ez/mdoKO2xGzZQS6lYkVPkESN40z36qH249okTnKbZ3DwO/XaXLOH011kfjCkMrI7duZMPv+l7AfINIHkASQD4EN9JXLjANAy2piuNc2rWFFmwION3/yVPHnql+MLSpRTMISG8Z40KCYCMl4hQwy9VigNAWJj5+M6d+dyUK0cvlXfesT9+8GA+X0FBVKAqV+bsdtIk+sVHR9PE8913vGf376fitGUL16AiI+lttmULzZNdujBytksX/r9lC1MbREZmTnOfOpWeRq5SbFy86NwUa8nu3Xw+bRd2z56lcrdgge/99DPZW7h//jndwDxZlEpPp5tWkyb2ppsdO/iDnjpFbeStt2jmOXfOvI/BwGmbpV1cnGjuPXu6XlCKjRV55BFZ2r+/vBEaKn88+ywfEstB6tVXxQCIAZCbSmXeHukP0tM5zW/UiKaA0qUpXJo0oSvmnbJweaeyZo1IeLj8V7eurMmVS857UivXkps3mSxu1ixqzKGhNPGVLk1NXhxo7t27c59OnawdBMaNo5AtW5a/aeHCItu3Zxzf3VT0e84cCrUGDZiH/umnacocOpTPjcFArb9WLQr7Bg347AQE8JyuzKQmfv6ZGrUvNXgTEjjD2LHD9X7du7OWrSeMGGFtpjKxdi1NSHeIQpN9hXtqKjUWTxMNmaagzjIEvvBChvYiIrx5Gza0FliTJlGrd5X467//+NC5uqnffZemGYAPAUAvBdMC282bXIxt25Z9eP55LvTcTlJT6Y1Rowa1F5OWlJLCvlatSpusJ0nRDAZOdXft4vH+ZOdOukI2b0676i+/2Pfp0iXul9mMjL5w+DDtzfPm2d+LN26IfPMNteKPPjLbhA0GCuOoKJoTu3bl37x5uUZUqpS5OPucOTQt/vorlZRixXg/Vq7M1Agmrlzh5yNHckYxcyaF5EMP8VWoEO3kNWtyn9BQHmNJejqFfY0azJeUlsa+VqtGG33fvpwZuMq7Y2LGDDo7eKsgfPklTUKu2LePg4enyePi4/l9HRVN79uXDhd3ANlXuP/yi+cpQpOS3NcP3bOHi0gmDxuDgfbA1avN+6SkcGGqRw/HGteOHfSecZZ3Q4SZGgMDmfTr7Fku8Jw9y+lrQABvntq1qQF98w3zwMfFceoaF2dux2CgWWnZsqzJeGjLkCH09HA2OCYmUihYDpCOWLGCU/ycOekxki8fbZmWD3VCAjXTH3/0TrMdMYK/YWwsTWNTplCjfOghtnnpEu29BQrQtlqgAH9LWyG/fDk9SG5l/MDy5WaXvPvu4yyxQAFezzfeoMC01U4XLuS9VL487++AACoN+fPzfe7cnFl17kzF5upV3lOff85n4STuXi0AACAASURBVK23eG2ef57a+dixFPIzZ5oT0j3/PG37lguNJkaMoPJh+RutWUOBbppNjx7Ngd/doG8w8Ht4m9O9bl33PviDB3PA9IZ+/RybzLZv5yDrj8yumcRvwh1AbgCbAPwLYDeA4Q72yQVgHoCDADYCKOmuXZ+F+yuveF4y7ttvM0qguaRcOesCA599Ri3JkuvXOcUrWJDTvBkz6OrYogVNO1995bz93bu5iOTsZvzhB/4UTz1lr8GcPcv+LVxITahqVS5WNWzI82Y206QrLl+mJuMuF72pYLOzsmxbtnBRr0IFCu5Vqzj9zZnT7Mq3dCm1xubN6SkUHk7B546FCymMbL2c0tKoXfbsSQ3zxRfNGujly1zge/BBDupXrpgX+vLl499evdyfO7McOcLvHB5O+/b999NTKiiI/QkJcW4KWLDAPPMLDOQrb15+FhLC6121Ku+5IkU4G+zbl4PBsGEc3GrXphmxY0eaVqKi6O31ySe01ffoYS/MTEF+tumvu3WzzuNkMLh2o9yzh79JZCQHJKU4KFWsyHZcza5SUuho4E65qVzZ86pSJr77zrnHW9myd0QeJ38KdwUgv/F9DqPwrmuzz/MAvjC+fxLAPHft+izcH3rIecV3W1q1sq+Ic+YMgzOGDqVgGT2aAsVSOF+5wpvH0ULN0aN8OHr14sPy9dfui0K0aEEB7YwjR3iDt23rePvcufzeLVpwamgaADZsoFC1nTb7i/Hj7Qc5Z3TsyKmyI0xRj7auan37UgidP0+hY8onLsI1DkfH2NKkCU0dIlysi4mhjTkhgdfFVH3IdtA0GKj9/fQT2yhSxCxQliyhsLGsgZoVvP46hVm+fObBKT2dGnOOHPxtXa0rRUXx+uXOzRnfww+zvZw5ue2TT6iZ//ST+Zhz5yjIt2zhDKdSJc4YunXjyzR7MJVotGX2bA6+tuTNa2+SnD7dPiBw+XLey0WL8hk8eJDfPVcu/marVpn9z/v0cWwi+ecfzsBcceMGn2F3948t+/dzFu6ILl2sSyzeJrLELAMgL4BtAOrYfP4HgHrG90EALgJQrtryWbhXqODR6DllyhTZniOH/GSalq1dS1tsaKhIr16yuU0bGRUSIjseeYTT2eLFqembzDOFCvmvPmpUFH18jdjVu5w9m4tjxpvKbrE2KcmsmSUlWW9v08a5i1dmGTDAzs/XaerWTz6x+o5W5MiRUcDb6rvv3cuH+sMPOQDYtl+vnuvUEQYDr0tyMm3ahQvL648/Lj8CMrtGDe5TrFhGCL5d3z/7jP7nOXOKTJ9u3bfq1TOfxdMdDz5IE8xzz4mIeVG0Y5UqfDTDwqhMOOp7aqp53ebkSRn1yCPSLiBAhj/2GO+HGjW4TlS2rIjBYH3822+bzWgGA6Okv/2Wr/XrXdu/Y2P5suTGDUkLCrK/L3bv5vNq7P+Q0FBJDAnh/W675pI7t7UmfvYsla/ixe3L9i1b5v63OX48wwvNXbpiq+3x8c7z0g8ZkjX1AbzEr8IdQCCA7QCuAxjtYPsuADEW/x8CEO5gv/4AtgDYUrx4cd++2f33e1SjMSYmRjYA8njhwgyYiI7m4pRRy7VyRxw8mMEd9etTO752jQ+WJyv+nhATQxumEbt0wHPm8CYuXdq+byLUaIKCqOFdu2a9/bHH/Fub05KBA+3qsjpN3frRRxnpju3IlSsjEZvVd9++ndvGjKGJwLb96tWttXlbDAYOHNev005esKC0CAiQPwEZqJSpwxn50u36/tFHXFDPnVtk/Hjrvj3wAO+FrMTkpmgMEjK5IlYyxV+EhGTY/+36npbGgU0pkfh4675//TVNIi1a8L5KT7c+/pVXfPfdHjaMjgGWJCdLirHvVvfFzp0ZGvYrYWFyGJCGERH2bRoMHGAdpQaZNYtavmVKgD//dC/cT5ygUiXu0xVbbb96ldfdEa++ypn+bSarNPdQACsAVBIfhLvly2fNvVkzj+zMU6ZMkV9z55ZD1avT9mYzvbMarTt0oLZz8yYXPuvXpzbvr8WTzp2togTtNPczZ/iQGk0gdprGtGnMX9Oli8iAATJ94kSJiY6Wpc89R9NFVi2sTptmN612qgU1b07NzxH/+x9nTJcuyWuPPy5VAgKk7xNP0AyVNy8Hr2LFRObOlSlffikx0dGy/Jln6A3i7jdo1Soj57h8+638Fx4unyslT3XtyuuaPz/NDKmp1n1PTeXny5axH8HB8marVhIYECCzatbk72G5qJ4VxMXx3syZM8MdMR8gB/Lk4YBjEV/h8LqXLk3B98or0qN7dwkMDJT+nTvze1WpQiFfrZrI5MkZx897/322u2+fb33+6SeagGxIzpNHakZGWvdv3DjOljdskMSQEGkcEeFYez592nUVp8mTuehq0vZ37XJt5hSxssvbXbuTJ3ltBg0SefRROVOmjKzPmVP+q1uXprJixRzPXkxy4jaTZd4yAGIBvGbz2a0zy7z5pudax/DhfHBOn3a9X4kSZn/ytDRqG1Wq+NY/Rxw/TmFhUVTCinHj+FO0aWNv59+/n7OOv/+mTbh1ay7qFi/OG95datjMcP06z3X4sOv99u+nwHCWlG3XLmrYAQFsLzqaWmdAAItFizAnUJkynJlVqMCH1xM30BUrqKHZCqsbNyj4X36Zs5sOHcxeU/v3042uZUvatJOT6V2jFPsUGGhvesgsCQkMzd+502xHP3uWC4olSvC8hQrxOuXJw0E7f36ux2zdSkHUuTP/bt3Kz0NCqPlXqMB7oV07DqL16tH2npTE61KuHGdBzZtzu2kw9IXUVP5+pjrBJvr3t656lp5ursvaqRPXuZwxYoT7MnsPP8wZrqkP+fK5r0BWowa9eEysWMFrFBbGazl6NF1HV63iWsDUqTxPcDCv38SJZpu9wUChfwfEnfhzQbUwgFDj+zwAVgNobbPPQJsF1e/dteuzcP/rL2o7nvjFtm7NH3LbNuf7bNrEh8v0wN28aU536m36UlcMG0bh0bmz2dXu6FHebErRdNSmDbWxESP4APbt6/hhPHmSD21mMgt6yujRfEicmajOn+dA+NlnjrcnJ1O4NmpE7w2TZ0dkJL9vZKRZMCcksJ1x47y79jNn8jr17s1kbEOHcqbUrRt/z6QkuvSFh1NgFi5Mu7PtYtupU3zQ3WUOdcT27VQmhg+3zumfksLpfFgYBWzp0hTEJlPajh0c0EqW5H1dqRIHwC++4LXIl4/9HT6cC+vDhvH/fPm4/Y8/aFcvXZrtFyzImZKl91BaGpWDn3/2z+L75MlUgCyD/f75h9c8NZXP5osvUlCaPKmc5VI3xa3YDha2LFiQsS4jImzbXdTo0KE0u125wnujZEku+ruqC9u9Ox0Jli/n4F+xIj1u1q7ldb4Vz5wb/CncqwD4B8AOo/kl1vh5HIA2xve5Acw3ukJuAlDaXbs+C3eTH7qxLqVTjh7ljf7553x4jKXO7HjqKXOEqMFAs0zbtny58lv3hW++4QNgGcRUvLjZo8e0uPXSS+Z+uXNDzGoMBi4kFS/OnOMmIX/hApNLxcRwcc7ZYDtsGLMGmswrSUl80E37T5lC08HChRS+TZowG2bhwt65eZ4/z/7068epta2wuHmTLnb9+nHwsJ0hpabSvvvxxwzd9xSDgWs2UVHmIi+RkRTo6en0Inr8cfOAbjBQg4yJMXv5mD6bNIn3gmlgGzyYA2v16ty/QQP+rV6dL9MaR3o675vFixkkdiuIjaVSNGmS2QW2cWO6nzZvTk+ky5f5+w8Y4LydkSOdZz+Nj+eANGMGzxMWZr4nZs/mgrErjh+nkhYdzTUvd8W+z5/PMB+KCH8XU33kGjW8T2ecRWTfICYRPoSVKrl2cxo50ryI+cYbHIFtU3iuXs0H8dIlju49evCmTEigO1xWFe09fZrmlNstuL1h7VpqNblzc3E3Tx76kLvKR5+SQpuwK++m9HRqncHB1ula16/n4OzOpOYJ6ek00TRvTm23WTPOGkxaWFoatzdsSDNOVJR9Badjxyik69ZlRS4Tv/xCLdZSI758mSaJMWP43Rzdp2vWuF5TMNV/PXeOQmbvXs4qTLOcs2cdR4/eSpYvp7krf35q3wUK0Kz05JNm9+AHH3SeIGzSJB5nGzS2YwcHhLAw/lZPPcVZbNmynLE8+igHxqgo1+sie/dyTccy66orXn2V57Fl2jTe82PGeNZOFpO9hbvBwAftpZeca4wvvGDt6WEqYt28uchPP8lPb70lxwMDZaOpuIOpXJ5pcfLAgQzvFY0FBgOFlSdmsc2baWqwwOHCYNOm1Ixst/fpY36gjh+nMHj/fbqjzZvnuflkzRqRChVk6qRJEhMTI1MnTaJpxDQwLV0qUrWqTJ08WWJiYmT+sGH2JeOaNZNt//ufPFWokCTlzWt2k+3UKcNsZtX3qVN5DouQdavtBgMHBWf5x+fN4z3u6rq1bm3W/m8n165xDeDcOXq1lC/PILW1a/kMWQY8mQrcdOnCtYCDBzO+2/TPP+dxUVHU+G193EeN4mxl1iwK7IgIzigdORQkJrL9yZP511iE3Cnr1lERsXV//u8/KoDffce/7iwGt4DsLdxFqG1XqsQFVkd2sH79+MNakpRE00i9epICyFVAdubIQYFhG+Fo4Ser8ZE1a8wpkPftE3n1VVmWO7f8Dsj3efMy50l6OoNajB4YVm5p777L2UG7dhS2vXvTVj5kiDmKdcgQx4EulixZItKsmXXbTZuag+G+/16kXbuM7WWiozPSM2fwwAPStnBhCQLkVGCgeRb46KMZUZhW7S9eTBOKxUKinUteo0bOw+6/+orf3dmxIpxp3gFBNXZcvUoTXtmyvI5PP03T3aBBXHcpW5bbjQuiMTExkh+QzTlzcnbobEY+dizbMDF7Nl1pGze2N7MNHmwOwDtwgL/FqFGOZcXhw5xBWAZ7iXDwj47mQC3CWVrp0v5di/OB7C/cRWj3bdiQP65tOPTrr1uv3otY2dA2tWsnMdHRTgMbZNs2mnI0vnP6NKfqDz1ELevtt+X3gQOlR3i4rO7WjXZjk4dMVJRISkqGFjdt4kTOtMLDnS+CHTjA2Vt0tOtF83PnRIoUkT8GDJASli6kptD+U6dEihSRv/r0kfpFi8q++vXto4VnzpQb+fPL4aAgOVKtmllIjBhB4SU22vVTT9FkYeFKarX94kVeG2fpBbZupU3baLax09zT0qi1uvreIgzzHzSI5qSaNdknU7KvrCY9nYuYQ4ZwBvPpp5wl2QjYqZMny1+5c8te24R9tsTGmr2rTPz8Mz3iHn7YHGG8Zw/t5JZOAEePUoFo2NA6UGvLFgp2y7oQZ87QRBMRwXQZlnTrZt+HW8y9IdxFeJN++intcx068Mc+c4Y5NapX54i+cydXwE0ujp4UxzWttGt8Z/9+2ujbtXOcBdJgoHubUvxdGjWiB8T8+dSQ8uenrdsd8+dzSm07wFuyfj01xoAAmg1s84hv2sSgouLFqRE7crM7fJjCwFI4XbxIAfbGG+zr0aNULEqV4uyvaFH7XOJpaRT+zzzj+ns9+KBzt8Vp06yKddiRlEQt2BTmv2YNv+P06Wy3fPlbU0buiSfcF9SZNYszPFd52UVYK8GRGWr0aP4GUVG8f2wzvZpIS2NfTIVI6tXjADt0KAe80aPN7qTPPmvtDWRi1y6aZ9z1NQu5d4S7iWvXaIZp1oxT+NBQutzlykWbW8+enAJ7Yis2LQTu2ePfPt5LpKRQmL73HrVvy0VIE8eOccB96y1qSYMG0Ub/wANcCPMm0ObTTz1LEudMY01Ppz113jzXg4QjTp6kGTA8nLONfv34mQgFakQEzUDNm5vd6x5+2DwbOX6cJqh27bigt3o179OdO3ns8OFmW/C5c/w/IsJ5pHZaGu31nTo5X5eYMYNCKqu9a/76i+ZTV89d/fr2GrIte/fyOztSEhIT+czPmUOhHRBAc+3atfYpDZYsoaNFcDDvtZYt+du0aMHF9Fmz3PvPN2nivr9ZyL0n3C0xGDjdffddTou9zRU9bRofPo3vzJ1rvobr11MjrlePC6STJplT8H78MQVrixa0z3bpQmGQM6fn2T9FKMTCw32rXL9jBzXZ0qWZBqFgQWuPD3cYDJw9dOjA14IF5nvu5k3OIAEqGkFBnKmYFlp//53BSy+8wOM++YRav8lZ4MABmn2Cg/n9goOp8bv6nnPn8nu40y4//NB1zV9/YHJdduYxY+kf74qXXuJ6izNefplKwoYNVCoGDKAZypReOk8eKnyPPMJrn5m0zqNGOc+jdAu4t4W7iYQEPljeJNlfvZralztbpsY1jRpZ57xJTeWC1QsvMJpx3DizjfT77ykAgoOp8RYrRtts+fLeaUhDhmTksPGYGzdos69Th+6LHTtSiNaubZUPyCWxsRyQvvmGC5wVK5qjqDt14izE5A5qMHBACwig6bBQIesoShFqjhUrWi/wJSdTa/cky2Hjxp7lGzLlUDLNMrKK8eOdF7sZPNi9DfvwYQ64xiRqDtm7lzORSZOszV1pafyeiYn+qxi2bJl1QNUtRgt3E6dPc8HuxRddT7fS0vhwhodTsGh8JzWVJjEHmqNDl77GjSnIWreWqXXqyFmAeXe+/94uQZTLDH9//snFW2+YNYv2/rp1pXfXrhIYGChvtm7NgSY01CpH/Yzx46V20aLW5zblvD9zJqNv3370ET+7cIGzkUmTRMQmp1DlytQwLaoJWX23WbNoMvCWK1e4VmFz7Z1etx49/B+sZ8vNmzSXOipk3amTa3fOCxfkctGi8k5oqHPnBxHO/gIDqVF7WZ/WXdZIO06evK21jbVwt+TyZXOO6Gefpf/1pUt8cA8e5PS0eHFqa1u23Lp+ZVeuXnWaCMqhS190NO3vPXrIG0rJf6bshgcP2uXWdpnhb9OmDH95jxk+nIPLO+9YZ1YsU4b3hEUukWEFCsgXtufevJkRtrZ9q1qVazxKZbhqWrXfrx+jJy1y2Fgdv2OH+5zljjh61CpJncO2LXnlFd+LdXtDfLy5AIjl7KNFC+cFPYw+858FB7vM6phBcDCv6/jxrvc7coSmqyFDRNq2lRW5cskiQGbny8eBeONG12aiCxfs4yBuIZ4K9wDcC4SFAbNnA3v2AFFRQI8eQJkyQGQk8NBDwKFDwI8/Ahs3AjVr3u7e3v3kzw/cuAGkpNhtio2NRUxMDGJjY80flisHbNoEXLmC0rVrIwRA165d+Vm5cub9Dh3CzJYt0atQIcS9/rr9eePjgZAQ7/pavjxw+TLwxx/o1bkzAgMD8eZjjwEXLgDXrvF+MVLsgw/wZWSkfd+PHgVOncr4bqNfeAE4dgyoWhUICsL/2zvv8KiqrY2/m0CAgEAaiCRAaIoCUiLSpSsIRJEWCdIEQRDQi+VeETReRbmKiihKERWQIKEIKIgURfhEIDRBkCoYCAghlCQmk2T298c7w5RMOTOZRrJ/zzMPYWbPOWvOnFlnn7XftRaWLwcMnykoKIifbccOnn/79tk+Nvv2AXXquPZZAKBKFSAjA9DpLJ62edwB4MIF/j68TaVKwJYttCsmBpg2DTh3DrjtNuDGDdM4KYFffgGGDAFiY4FJkxDy7ru2bTdHr+c5V7Ei8M8/hV/X6YCvv+bv/f77gaQk2jRkCP4ZPRorwsNRMy4O2LMHGDmSx/6NN3h8rMnJAcqWLfIh8TpargDeePh05q4gv//OGGePHtRxf/ih/YJORcW8ip8zVq0ylSC4cIH6bmPCyNq1jGNHR8ubi5LGCpPt2ln2x332WVPbPlv8+ScX3MzllTk53F/r1lzI7NGDcfD77uMinhbefJPrA598wkf9+qZ6RfHxXNQzym8LChhjLlWK6fuRkYXbCaancw3ClsJIC507azv2V6/eDCn5lEOH2CsgNJR3Zs2a8VgPG2aZ5GSvHpS9bdaoweNvyDu4yc6dDM126MBQnxYZ4969vAsIC2PJCnP568aNVMz4CaiwjOImOh0ldrffTv3v2rU8yQcM4MlrrcP2BMuXF67Nk5HBJLIFCwqXTu3ViwqZPn34QwwO5g980iSGNvr25ZrIiBF87rPP6FDLlKHjzMqiU7ZVnnjbNv4YIyNZ2jcykrF540LmkSOMu9esSYWFsRSFKy3a1q2jrnzwYJaRNZKXx8U3IbjdsmUZGzYWodq6les8I0awHn5iIp3USy+5vwCYnMzwlDP7X32V54C/uH6d37FRNbVgAY+HO5UXx4+nXt285EVBARU21aq5X6bh2DEmP7Vvb1p4fvNN+41pfIBy7goTY8ZwQctW2vTu3czmc7X7vDN0OjppY5u+Dz6gc+vdmzkHkZEsf5yVRb22cQablMR070qVmG4uBH/0NWrw79q1OdbY3OShhyhze+MNFv+yZv16jl+yxDRjy82lI42MNC2e6/W8SKxdS925p/nqKzqIuDjLuw0pqYKZPp3rQhMmFH3dp6CAC5W9e9vO7NXrmcwTFeWdz+oqxhm1u9y4wbuAs2f53YaHM16fkMBjbisZyRXy85ntHhPDyUPbtkyW9BPKuSvIqVM82c1UH4VYutR+2dWicPIkwymDBpl+GEb++YfOuWtXzqysO9x89hk14R060KHXrm1ZAfH77/n6K69w9l6pUmGpXF4eHZi9Yk+bNnG27olU/MxMXoQWLChcyOr552n/lCmUgUZEeL8AVW4u79bCwznL/OEH7vPDDymzbNLEvZwAb5CUxDsqWwlKWpg+3UJ1JJ99lovcnTtrz1XQwqxZPJ/uuMO5Lt+LKOeuIFOm8GR3hE5HaZc3uswYmzXcdhvDKUlJUq5YwbBE/foMUSxZUvh96ek8PStX5r+2UuUTEhi+KV/+Zr9MC775xnm515YtXasbb4u8PNZu6dmTMfs2bUw//gMHaJtR0y8l1xMaNvSc7toRp05xHaJzZ4bJEhJ4h+SLfWslP5/OOSHB9ZDMqlU8d80nDrNn87zyxsWrQQPmNPgRrc69ZKhlSjInTwLNm1s8NW/ePERHR2PevHl8okwZoEkTjvU0NWpQdbJ7N1CxIk7973/4fvBgHFmzBliwAOjWjaoFKYEDB4B16/j47jugVCngkUeoTGjdGhg0CAsqV8ZsIbC7XDlg40aqIF5+GcjMLLzv338H2ra1/7kBvr5tG/Dcc0D37kD79kDv3sDs2VTfaOHIEeDSJcyLi0PNgwdx/dgx4I8/+NrmzcCjjwKhoab9p6VRjZOWVoQDq5GYGOCtt2jHzz8DixYBnToBQnh/31oJCqKaLTUVGDCAqhdnSMnzZ8wYYM0afk4ASE+nymXQIODFFznOU6xYwe3pdDcVUAGNliuANx5q5u4F8vM5w92xg/Wpjx6lAsGqcJNNzXPr1oVVG54iPJzaYlv7btKEYYumTRm6Mc5+77iDM/b+/TkzT0uTcuFC+RwgxwOyF8D4eKlSXBi0pTueNcui+UKhfR8+zEXmSpW48LZhA3t+JidzobFKFSa/Oasbn5YmZWiofLhqVXkfIDOEMMV5Fy/mZzLbf6Pq1blPZ52BSho5OTxfIyNZfM3W4nhmpql7V8OGhesPPf88k6X++Yd5AoYEsiJz7BjDhzt2MIvdU+E8N4AKy5QQLl+m0qBtW9bQqFuXIYL776ezLFeOjmTmzJuhgUIZeSdPMg7syfikOc88c7MOt8W+V6+meqRLF8bQzW/J8/IYS4+J4WlqqDNTu3ZtCUDWrlWL5QJiYxne6d698H6NrRYNC8kW+/7lFzqRkBD7t+/nz9PJt2vn3BEnJckboaHyXFCQ3DR6tOn5rCzGaT/4QM7/6CMZW726PH3vvVzkVtjm+HGW3A0P53ndvz8Xm3v04PfZp4+pF4A52dk8j42F306c4EL8p58WzZ6jR+nMjXXdpWQ4b82aom3XTZRzv9UoKOBJtHgxsyanTGEJ0vXrC3eHkZL69HHjGJN+4gkumNlquXbxIp3YAw9wJjppkqWjKihgsa7nn/faR5MXL9IBP/kk9cOnTvGzlS5N9Yi9OOugQfyBt2jBU/XBB6m6mTaNF7HgYMbwhbDfys8oTzRfADtzhjP2Dh147BxRUEBZpFlddpc5epQXH+PC76RJrsksSyrZ2Tyvk5Iog1271nEJ6EWLqJ4y5/hxnivDhrneklCvp0OPiOACvzlffOFeeQgPoJz7rcKVK0z/jomhoqJfP2rRX3uNC6GdOtGBd+/Oma5ezzKqNWvSWdpy/Nbs3EkH/+qrVKjUqcPww+HDLJTVvn3RZ+1ZWVSuREVxNt66tWXRr0uX+LkaNOCY2FiOcbSAlpHBhdgGDbgoKwRPWYB3JJMm0cFbN9YwJzOTP/iGDZmM8u23dOphYZRO2mrRZo1Ox+9HSx8AR+Tm+u1WvkQwfHjh7mtSUik2dizPu/fes1zctkV+PhfZO3VivsCBA4XHGGv4+OH7VM79VmDlSq70DxlSuHmEOcb2gE2bUsYWEcHbUlc4dIgz9CpVmAEYFMS/p04tumPX6Xhn0L8/95OVZcowtdWoQa/n59i82fm2//qLMXeA/06ezB9qZCSfM7Tnc4heTx3/sGGUXpYty9mgK4qRGTP4fkXg4qwRz86dphpTvXpxsrNiBUOC69ZxjeaJJzhxuu8+ztYdZbPWrcsJko9Rzj2QKSjgrLNePcfd261JTuZMNizM/aSjv//mSZ6UxAUi82xKd1m2jE7WehZ+7JhFzPsmu3YxTKPFuZ44wYtZUhJn6eHhnIH17ctYaqtWrttqfeuuhUuXOFNzV4ut8C55ebxoa2ma/vffzKB+6SWG27p25TkxejTPqYMHte2zX7/C+Rk+QKtzV1JIfzB5Mosj7d4NtGun7T3nzwNjx1LSlpwM9O/P4lOuEhnJwkkDB1LaNXw48Pffrm/HnDVruJ1SpSzlhvXrU2L500+W48+cARo1spTj6XTAhg3YNnQoJoeFYf2kSUBeHrB+PeWQAwcC3btjMKuEJQAAH+dJREFU48CBiAYw76GHgBEjgMOHWfhLK+fOAXfeafMlm1JJIxERQEgIpXaKwCMrCwgOBsqVcz42MhLo1w+YPh345hvMGzAA0YcOYV5sLDB6NNC4sbZ9hoVpl8v6AeXcfc3q1XSG69ezgp8WpKSed+xY4L77qFP+/HMgPr5oJ1fbtsDQocC4ce5vA2BFvlI8lRITE5GamorExES+FhRkW2tsdOwXLwJTpwK1agH//S/SkpNxd0YGIubMAWrXBr791lRdUggsW7bMtH0h+LC1fXtIaVfjXch2ezYHGrm5rCK5Ywdw7Jhntd23CkX4zE6/d3u4eu75GOXcfcnVq8DTTwMLFzousyolHaaRXbs4Q335ZdNzDz8M9OgB2Cp96wqvvQbs3Ans3+/+Nh5+GPjyS0BKy9Kyf/4J7N0LdOhgOT46mglGBw6wrOvff/OOZPt2XH//fUyLisL+2bOB779nSdglS4CUFODIETw8Zoxp+xs2sGRveLh2W6tXB44ft/mS3bK4AMvoZmZythYonD8PvPQSj+cTTwDPPw907syy1fPn886npFChAi9yNspMO8Ph9+6IjAzXS0z7Ei2xG288SmTMfeZMLuhYk5HBmh/9+lGVUaoUlSHlylFV0qABlTHWMe2MDC4OnT9fNLtef53xRnfJzaWufuhQxsizsihhu+suU/VDI1evsqxv7dqMn2uJWbZqxXh7vXqmOP3u3Yy9u1rAydhaztUemjNnUlJpD72eMr0zZ3yT2r9/PxO9nnnGsmxEQQEX2zt0YBzZW7kLgUijRr5ttlO/vv0m5V4EakE1wNDreTLs2GF6Lj2dyo/KlalkWbyYGaZ5eRx/4wazRsuWZRnTOnVYmMrceTz1FMvEFoW0NOqvtdS5tseRI7TRKFUMDqYzNErFtm9nslKFClxkDQ7mgq6WWiLZ2bQvKIiLXy1a0LEtXuyerePGUZaplfz8wt+dOV98wYtVxYr8fDEx1Fx7iwsXmJyTlGR/TF4e8wQGDvSeHYHG0KFFT1jSytWr/K79UEDMY84dQDSArQB+B3AYwEQbYzoCuAZgv+Ex1dl2S5xzP3GCDsnomNet4//Hj3fcLGHnTkog9XqWFGjenJp3Y23pzZu1yQGdcdddtvW8Wjh3jg0vXnyRdxE6He1u354/uM2bKV384gsm75w5Q9VP+fKc8Tua6er1lD+2aEEHv3gxNfpFuRCdPMkyx5s2OR+r17NRdvfutu18911enKtVk/Lf/6YCo2pVPvfee+7b6IipU7VluObkUGrrh9mlX1i40HbZZ2+wZAnLaPsBTzr36gCaG/6+DcAxAHfLws59nZYdGh8B6dwzMuiIPvyQZUTfeot/b93quLm2FpYtMyXbfPIJZ15ayr5+9JGUI0ea/q/TUZ9bsyalhp5Kphg8uHAWnlaefprp4tZkZXEWW6+eZfhk2jRe1I4eZZZqly6UhJo7T72eTrxXL6Z6X7rEu5TXX3fPRmt+/NF0wbF37NLTmRgTG2v7+8/IYPmCJk0s5Z43bjBEUKGC5ztdGSt42svItWbqVN4dlgSysnhXaKhj5FXatqVG3g9ode5OF1SllGlSyr2Gv28AOAKghifi/QHBuXPAq69yYS46msqNw4e5WHLlCnDoEPCf/7C64V13Af/9r+2+is44dYr7SEpi1bpt24AHHrA51EKSd+4cVSNGypRh/8kpU1hRMTeXEr2LF22/Xyu1a7MqnzssWQJMnFh43yEhQFwcbevV6+br62bMwGadjpLEJ5+kfG3kSKBRI5yKjcW6kBBk1KgBPPUUKzVu3UopYrt2rMDoCR54gAu2H3/MfplvvsmqiXv2cKF2+HD22S1VCvjxR6By5cLbWLOGn/H114EKFTBkyBCULl0aQ8aOBRITgfLlOcaTnDrFfd5zj8XTdr/zuDhg+3bP2hCohIRwYfn99727n127KBbo08e7+ykqWq4A0jRDrw3gLIBKVs93BJAO4ACA9QDusfP+0QD2ANhTs2ZNH1zjHHDmDDMqQ0M580xJcTz7zctj8s3o0VzEfPzxm13tNfH665ytRkRIuW+fw6EW1QtfeMHUj9OaF19kHeyoKItGFXY73Tvi1VdZPsBVCgq4+Pvbb1JOmCB/DQ6WuwH5U9myzPh7/nkunJrZtgaQw43PTZ/OMXq9lDt2yLFhYXIAIPtERhYOgyQnS/noo67b6IyUFH6vbdqwn2enTqx946y0wwcfcKZoCGcFBQVJADIoKIg1dMLCTJ2oPMW+fbxTsMLud370KO+cSgoXLjBEVtRSEfbIzeVdmTfXVJwATycxCSEqAlgBYJKU8rrVy3sB1JJS3gvgQwCr7VxI5kopY6WUsZGRkVp37VmkBObNo1yscWMm1Hz0EWueBwXZf1/p0tSYf/opr9p16gBNm96UADqlUiXWKZ8wge9zgIU0KyTEdq1ygDLGP/5gYo2ZJMstadeNG5STucr583zfAw8A166hSvPmOFu+PKrFxgI//MBZ1LVrwOXLN23LDQnBsJ49+f4ff2SikxBAmzZo9tZb+L+oKPR6443CuvKLFx1LSN2leXN+rzt2ULq5ZQslps7O0SZNKFldtw4AEB8fj6CgIMTHx1Ofr9dzjCepWpV3WDqdxdN2v/NTp4Bq1TxrQyBTrRrPuWHDtNWFd5XXXuNvf/Bgz2/b02i5AgAoA+B7AM9pHP8ngAhHY/wSc8/JYaGsFi08s8i0dy9nUQkJzhf4Fi5kVUBX09e//tpUkTAri4uqycl8fuNGLugFB7tlvgXdurHqniucPs02eu3bU7ZZty5VKG+/zc7xFSvyM0dE8DWjsiApibPjNWsYP9aSMi6lX+OcNtHrOSuuWJFrKvn5fCQl8bkGDbwji+zYUXvD5759PVfT/FZBr+dCfs+enq2+OX8+hQNFlR4XEXhwQVUA+BLA+w7G3A5AGP5uCYZuhKPt+ty55+SwZGy/fp79wrOzuUIfF+fYwY8aRSfsaoOGQ4cYBmrZkuqS5s0ZmujXjz/y8uUZFhkxwnXtthG9nqETV07a/Hzenj79NJ33ww9T/TNlCsukJiQwLPHII1Ti3HYbLwSffspb2nLluE9HBdPM2b+f4Sc/9q60yZEjVMeEhZmkkGFhDA14o22hlKyL0qyZcw373r0MOzrqn1tc0ek4keve3XkVSGfo9ZxERUd77zt1AU8693YAJICDMEkdewIYA2CMYcx4UCZ5AMBOAG2cbdenzl2vZ4y8b1/vOIfcXNZ2Nuv6U4i6dakKcUWHu3w5nURICJNobF2U2rblLLhlS/6QX3rJ9YvXTz+xJK4rrF1LFUnVqibVz/79lC2OGEF709N57J96iusblSuz5PBjj3FW1aqVtll7djbj4dYJUYHC1atcX+jWjY8PP/S8SsacggIex4cesu+4U1J4V7R8uffsCHTy8ligr0YN1+9KjZw5w+80NtY3KhwNeMy5e+vhU+e+dClbbmm9/XeHzEw6cFsZk1eucFa3fj1nsc5CMzk5nPneeSfDMDNm8DbTmu3bOZtdulTK3r25mPToo/ystlqU2WPgQDonVzBezLQ0LEhL491Hx46mZtjGJiFdujieWaWn8+I1eLDrzZOLMzod75rCw6WcOJEX6JQUlpHu1YvPJyf728rAYOtWJgD26MH8Ei2y4aNHeVzDw9kFLIDuGJVzN+Lt1XNzfvqJoYn0dMvn/+//WB9ar+cPz5EqRafjmEceMd12X7rEW33zfpHZ2XT+y5dL+fvvJkWEXs/u79HR2hz8wYM8gV3R8ev1jKV368bYvxbi4qjXN6+Jnp/PmVWVKtRi79vHdYWsLIYUxozha//6l2pyYY/Tp3m31rq1lPfeywze+fO1NSEpSWRnc90rNpa/jccfZ6hl/Xr+bjdt4uvjx/MuuGpVJqWZqdACBeXcjYwdazvBxpv7mzzZ8rnNmzlrlZLyyWrVOIOwxXPPyTONGsnaNWqYepxKyVv91q3p5AoK2ODDmFp+5gxn8ObMmsWYuKMQjU7HGL55b0gtZGdz/aBlS164zEhISJBBQUEyISHB8j3jxjFcY0vKmJrKxKY6dRiLNy7Ovvaaa3JThUILhw8zYe/pp3kxbNeOd4eDBjH0t3VrQLdBVM5dSsY9Q0N96yCMzSXMF7u2b7dsKmFszmyteNi+Xcrq1WXj6tULa5YLCtjtaPJkhmw6dDDNzv74g47RHL2es397NVT0es6aH3zQdUVHQQHrvHTuXKhJsIXW25wBA+jchwxxbV+BzvXrVKO0aME1hcqV+fecOSVzIVPhdZRzl5Kp+/36eX8/1vToIeXnn5v+f/EiwwvmTjQlhbPTwYNNyTL33y/l0qVy7ty5MioqynLmLiXj+WXKcEZunu6+ejWdtDVpabYvbno9E6AaN3ZfSdCuHcMsAwbw/3l5Uq5aJX+qU0d+LYT8tmFDU4p8RgadXs+etntc3qps3MgLed++rIJ55QofGzfyDiUiQlv9GoXCBZRzl5IzS3dXyYtCUhIXOM2JiuKs3pzMTDbBrlKFUsLISEtHnJ/PePq8eQyB1KrFqpDNm1P7bpQuvvKKlP/5j21bxo5leMNIaiqd7P33M5bvLl99xTuJ0FDG/WNiqGj5+GN+/ilTqNbo39/Uziw01HUpaKCyZQu/r23b7I8x1q/RUkNIodCIcu56PZ2msxRyb3D6NBdWzRk2jKoXW1y+zESgOnVoc5UqXNApX57PDRzI8IdxUTEnh86zalWm7995J52NLVJSuNh64YKUb7xBZ5OYWLSqilJS8VOzJhsKlyplu0RCVhb12GXL8m7mueeKts9AIS+Pi3I//OB87IYNvCirBWGFh9Dq3Et7I+s1IDh5kl18/FHmoFYtICcHSEtj5x+ALfIGDQL+9a+bLeluEh7OVPV589hC78oVdtGpWJEPa8qWZbGq/v2Bf/+bnYVmzWJHpebN+ZmlBC5dYp/W06dZpGvAAKb733130T9jcDDwzTdA69bspjRzJrsp9e/PUggnTgALFtCWMmV4LFatKvp+A4F164CoKKBrV+djH3wQuP124LvvgN69vW+bQmGg+Dr348eBhg39s28huO9jx0zO/b77gDvuYA2TsWMLv+fAAaBZM75Xa9u4e+5h67533mFNjZQUYMYMVrQUgnVYmjVjdcN33vG8c6lblzV30tP52WrVYqXFvDzWm2nXjhUWGzYEatbkRak4sGgRMGqU9vGjR7MGkXLuCh9SfHuoZme7VwjLgFtlc82pUAH45x/T/4XgzHzqVM6kzcnPp70ai2IZbdsZH89iZxMnAo8/Drz7LmfPe/fS0W/aBPzvf3Su+fnufQ5HnD3LMsnHjvGicvkybvz4I9K//RZXt27ljHXnTuCTTziT9xd797K0cJMmLPp07710zvv2OX9vfj7vOHr3ZqG5u+4CNm7kNtPTLYbaPWcaNGDpZoXChxRf5x4UBBQUuP12tzuiGykoKFxlsmFD1mHv3ftmlUQADNNIqbmTemJiIu5NTUXdFSuAzz4rHObRYosnCA7mBaxUKdZdX70ad4eEIEKvR+MyZXhhqVuXY4KDPb9/Z/z2G9CqFfDoowxvdenCZt6dO/P/cXFAmzas32+LL79knft33wUGDmTd+pUr+dyxY0C9esC4cQzBwcE5k5fnn8+vKNEUX+deuTJj127idkd0IxkZths8TJhAp/LAAyzXC9A53n47yw87Q0os6toVn5cqhR0vvEAH44zTp9lsxNPUqcMLUkrKzadsHrfkZDpUX/LLL3TmXbsCMTHAsmWchUdE8N+kJB67Tp34+PVXy/dPn87yrqtXs9lFQgJn/nffze+udWs6+MuXeWHLzrZ/zmzd6vnSvwqFM7Ssunrj4XW1zOXL7Lnpj3okubks9mWuRbdmzhym/c+YwZo3vXs7rwVy5gzHNW7MsgFayMyk6sZbGXfTp1PGae84nzrlu9ZnRs6cYRbw66+bWulZ1/PJzWUuQkQEx1WvbqqquWQJpZ32etv+9hvVUDqdKVv4scdsj83NlfL225kVqVB4AJR4KaSU7EjvjxKde/dKec89zsedPEnNeWQkyxP07GmpA9fr6XBWrmS2aVgY0/RdqQm/ciVllt4iJ4ep2/37WzpwvZ5SwVq1mEzmS559lp2VIiKY9euIbdt4/EeNYterggLKT3fscPy+Bx+kfl9Kfh/R0fzerXnhBW3F1RQKjWh17sVXLQMw3rplCxe0fMnWrdy3M+rUYcee48eB994D5s5lp53QUMZor1+nGqVFC4ZyFi2yLY10xJw57EXqLcqWpczvlVcoibznHiAsjHHssmWp0unXz3v7tyY7m7HyQYO4yNy2rePx7dtTvlm6NLBwIcdXqcKwiyMWLQI6dGC8fdo0HuM5c/gdAlQxTZvGRe1t2zzz2RQKV9ByBfDGwycz9++/l7JpU+90w7FHQYGU9es7nzHaYtgwzjrPnuWs/uLFotm+dSszY31VBCkri1X2kpNZhdOXx91IUhKrVVarxkYaWjCGWbp0YX38Tz7R9r70dFYXNPbULVeOBd5GjuRzgwcXvVGEQmEFPN1D9Zaka1f2H92503f73LKFXe/btHH9vTNmAEuXskdmnTqcxVv3EdVKZiYwciRliL7Sl4eEAA89BDz2GHX97tpeFFJTuYCq11O2aMaQIUNQunRpDBkyxPI9jRpR0RMTw56wdpK8Ckkdw8KooDlyxNQXd/duJowdPQosXuydnq8KhRa0XAG88fBZyd/Zs1kDxRcLq/n5rNny2Wfub2P1anaOOX7c/W3k5DDOO2qU+9u4VXn7bXZ+ql690Et2K1ZKaYq716xpN94eFRVVuFqnOdWqqRLFCq8DNXM3MGYMkJvLeKi3efddzl6HDnV/G3FxjNV27Ajs2OH6+y9coJY7JAT46CP37bhViYykPDEri8fCjPj4eAQFBSE+Pt7yPefOATod3xcWxuQsGziUx2ZmAteu8f0KRQBgbGrtc2JjY+WePXt8s7OjR5kK//PP3itJsH8/0K0bsGsXb++Lypo1vDANGgS8/DKdxsmT1JSfO8ewQ2gowwGNGzMEsmQJ8OKLzMZ87TUuEpY0Ll3iAnpcHHXsU6Y4f8+rr9Khr1oFvPkma+Zs2ODafufO5cLy6tVuma1QaEUIkSKljHU2rvjP3AHGXmfOZBEn69R/T/DHH5wtf/yxZxw7APTpAxw8yIJbUVFUybRqBcyfz2Sn8+fpgB57jAXSKldm8bBvvwXeeKNkOnaAM/devVifZ/Zs54lhp0/zrq5KFWayDh/O0gKulEuQkt/9uHFFs12h8CAlxwM88QRv1Tt0ANauNS2AFZVdu4BHHuGMr39/z2wT4Mx8yRLghx+A+Hg6q5MngT17uGgrJR1S06a8EJQpAyxfTgf/wQcleyHv2WeBHj24oNy5M2fijRoVHnfwIL+7kSNZwXLjRqBcOWYRjxzJ/2tZjH7nHWYZd+ni+c+iULiLlsC8Nx4+W1C1ZulSU1ZiUWqaG2uqR0ZKuWqV5+yTku3ZunWjLO/YMdtjbMkMMzPZ4DcqSsr9+z1r063GF19Q3jh1Kr+jbt3YAHntWi54d+3KevjTpnHc4sWm9xYUsMNU167sImUPvV7Kt95iopYxu1Wh8DLQuKBaMmLu1vz1F6sCXrzIeuiPPKK9sFNuLrBiBWuPxMSwhK+xrK8nyM5m+OiuuyhjdKfg1/LlwPjxTKBp3Nhztt1qrFjBcrutW1PeePYsE8MqVWJ54kOHKJNdsIDngDkFBay9v2gRMHgw1z8aNuTaRkYGE6WM4ZwVK7xTu0ehsIHWmHvJm7kb0euZbNOxI2VzL77IHqW2+o2ePcvZ+eTJnO117cqx3kjSGT2aXdiLKt1cvJgdmIxNtEsqmZlSzp3LZLagINYbCgpih6h58xzX/5GSdWpefpnnSOnSTFQqV07K+Hgpf/7ZP4laihIN1MzdBQ4fZpXA3bupRtHruUgpJXDjBuPZLVowMefxx5mk4g02bQJGjGCpWlsVJV1l8GAuML7/ftG3VRzIy6Nk8bbb3Ftw1um4jZAQ/yRoKRTQPnNXzt0aaWhPl5XFH3DFilzM9MWPuVUr4IUXgL59PbO99HSgfn2GH+64wzPbVCgUfkVJId1FCKb9x8SwKUNEhG8ce0oKk27i4jy3zfBw6uTd7SalUChuWZw6dyFEtBBiqxDidyHEYSHERBtjhBBilhDihBDioBCiuXfMLcYkJwNDhni+Y9KwYVxgVSgUJQotM/d8AP+SUt4NoBWAcUII68pKPQDUNzxGA/BBrn8xY88ebWWCXaVpU+DUKYaZFApFicGpc5dSpkkp9xr+vgHgCABr3VccgC8Ni7k7AVQRQnhQH1gCOHSIjZs9TXAw0/GPHPH8thUKRcDiUsxdCFEbQDMAVg0nUQPAX2b/T0XhC4DCEVlZ1F9roFDpWWdUqkSViEKhKDFodu5CiIoAVgCYJKW87s7OhBCjhRB7hBB7Ll265M4mii9lyjBBSgOJiYlITU1FYmKitm3rdNqTtBQKRbFAk3MXQpQBHfsSKeVKG0POAYg2+3+U4TkLpJRzpZSxUsrYyMhId+wtvtSrxwJkGnBYetYaKbndevWKaKBCobiV0KKWEQAWADgipZxpZ9gaAE8YVDOtAFyTUqZ50M7iT4sWTKLSwKhRo/DXX39h1KhRzgefPMmknapVi2igQqG4ldAyc28LYAiAzkKI/YZHTyHEGCHEGMOY7wCcAnACwDwAT3vH3GLMgw9SDulpli9nnXmFQlGiUBmqgUJ+PhOnPFmOOD8fqFsXWLmSdwYKheKWR2Wo3mqULg1MngxMnMjaNp7gvfdYfkA5doWixKGceyAxfjwLU82aVfRtHT4MvP22Kj2gUJRQSk4npluBoCDWCW/fnjXiBw50bzsnTrAT0fvve67tn0KhuKVQM/dAo1494PvvgeeeY5PrvDzX3r9uHVsJTpkCJCR4x0aFQhHwKOceiDRpwt6su3YBLVuyB2h+vuP3HDjAWvMTJgBffcUORAqFosSiwjKBSo0anIUvX87Y+TPPUC7ZogVLEZcqBVy+DOzdC/z8M3D+PPDUU8DcuaxBr1AoSjRKCnmr8NtvdOIpKUBqKnt8hoYCzZpxdt+xo3vdhRQKxS2FVimk8ga3Co0bl+xm1wqFwiVUzF2hUCiKIcq5KxQKRTFEOXeFQqEohijnrlAoFMUQ5dwVCoWiGKKcu0KhUBRDlHNXKBSKYojfkpiEEJcAnAEQAeCyX4xwjrLNPZRt7hPI9inb3MPTttWSUjrtU+o3537TACH2aMm28gfKNvdQtrlPINunbHMPf9mmwjIKhUJRDFHOXaFQKIohgeDc5/rbAAco29xD2eY+gWyfss09/GKb32PuCoVCofA8gTBzVygUCoWHUc5doVAoiiE+c+5CiCAhxD4hxDobr5UVQiwTQpwQQvwqhKjtK7s02DZMCHFJCLHf8HjSx7b9KYT4zbDvQt1NBJllOHYHhRDNA8i2jkKIa2bHbqoPbasihEgWQhwVQhwRQrS2et2fx82Zbf48bnea7Xe/EOK6EGKS1Ri/HDuNtvnz2D0rhDgshDgkhFgqhChn9bpP/Zwvm3VMBHAEQCUbr40EkCGlrCeEGATgbQADA8Q2AFgmpRzvQ3us6SSltJcE0QNAfcPjfgBzDP/6Cke2AcDPUspePrPGxAcANkgp+wkhggGEWL3uz+PmzDbAT8dNSvkHgKYAJz0AzgFYZTXML8dOo22AH46dEKIGgAkA7pZS/iOE+BrAIACfmw3zqZ/zycxdCBEF4GEA8+0MiQPwheHvZABdhBAiQGwLdOIAfCnJTgBVhBDV/W2UPxFCVAbQAcACAJBS6qSUV62G+eW4abQtUOgC4KSU8ozV84FwztmzzZ+UBlBeCFEavGCft3rdp37OV2GZ9wG8AEBv5/UaAP4CACllPoBrAMJ9Y5pT2wDgMcPtZ7IQItpHdhmRADYKIVKEEKNtvH7z2BlINTznC5zZBgCthRAHhBDrhRD3+MiuGACXACw0hNvmCyEqWI3x13HTYhvgn+NmzSAAS208789zzog92wA/HDsp5TkA7wA4CyANwDUp5UarYT71c1537kKIXgD+llKmeHtfrqLRtrUAakspmwD4AaYrr69oJ6VsDt4KjxNCdPDx/h3hzLa9YB2MewF8CGC1j+wqDaA5gDlSymYAsgC85KN9O0OLbf46bjcxhIv6AFju6307w4ltfjl2QohQcGYeA+AOABWEEAm+2Lc9fDFzbwugjxDiTwBJADoLIRZbjTkHIBoADLc0lQGkB4JtUsp0KWWu4b/zAbTwgV3m+z9n+PdvML7Y0mrIzWNnIMrwnN9tk1Jel1JmGv7+DkAZIUSED0xLBZAqpfzV8P9k0KGa46/j5tQ2Px43c3oA2CulvGjjNb+dcwbs2ubHY9cVwGkp5SUpZR6AlQDaWI3xqZ/zunOXUv5bShklpawN3kptkVJaX9HWABhq+LufYYzXs6u02GYVS+wDLrz6BCFEBSHEbca/AXQHcMhq2BoATxgUDK3A28G0QLBNCHG7MaYohGgJnm9ev2hLKS8A+EsIcafhqS4Afrca5pfjpsU2fx03K+JhP+zhl2Nnhl3b/HjszgJoJYQIMey/Cwr7Cp/6OV+qZSwQQiQC2COlXAMuLi0SQpwAcAV0tH7DyrYJQog+APINtg3zoSnVAKwynKulAXwlpdwghBgDAFLKTwB8B6AngBMAsgEMDyDb+gEYK4TIB/APgEG+uGgbeAbAEsMt/CkAwwPkuGmxzZ/HzXix7gbgKbPnAuLYabDNL8dOSvmrECIZDAvlA9gHYK4//ZwqP6BQKBTFEJWhqlAoFMUQ5dwVCoWiGKKcu0KhUBRDlHNXKBSKYohy7gqFQlEMUc5doVAoiiHKuSsUCkUx5P8BcvLEtmd9Kx4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(\"Local Outlier Factor (LOF)\")\n",
    "plt.scatter(X[:, 0], X[:, 1], color='k', s=3., label='Data points')\n",
    "# plot circles with radius proportional to the outlier scores\n",
    "radius = (X_scores.max() - X_scores) / (X_scores.max() - X_scores.min())\n",
    "plt.scatter(X[:, 0], X[:, 1], s=1000 * radius, edgecolors='r',\n",
    "            facecolors='none', label='Outlier scores')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(3,1,figsize=(9,6))\n",
    "for i in set(y):\n",
    "    temp = X[y == i]\n",
    "    y_pred = clf.fit_predict(temp)\n",
    "    temp_scores = clf.negative_outlier_factor_\n",
    "    ax[i].plot(range(len(temp_scores)),sorted(-temp_scores))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(1,3,figsize=(15,5))\n",
    "for i in set(y):\n",
    "    temp = X[y == i]\n",
    "    data = []\n",
    "    for j in range(X.shape[1]):\n",
    "        data.append(temp[:,j])\n",
    "    bp = ax[i].boxplot(data)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd0AAAEyCAYAAAC/Lwo5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAEW1JREFUeJzt3VGIped9HvDn39GatR3Lnq2HElneri+CmbAQOxmMW4vA2nVxGuPc9ELCLjQMbC/SrV1S6qQLXeliKYgQ2otSEB43hahjWseGYNw0hmxwF1LHs7KSSFoHUtdK1nKrNTut7LSKV8q/FzsWkmp7zuzOec/MN78fHDRz9jvne84czT77vef93q+6OwDA/P2VRQcAgKNC6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYBClCwCDKF0AGOSueTzpm9/85j516tQ8nhoADpwrV658q7tXdttuLqV76tSpbG1tzeOpAeDAqaqnZ9nO8DIADKJ0AWAQpQsAgyhdABhkptKtqn9cVU9W1RNVtVlVx+cdDACmZtfSraq3JPlHSda6+3SSpST3zzsYAEzNrMPLdyV5bVXdleR1SZ6ZXyQAmKZdS7e7v5HkV5L8aZJvJvnf3f3br96uqs5W1VZVbV2/fn3/kwLAITfL8PJykp9L8rYk9yR5fVV95NXbdfcj3b3W3WsrK7suygEAR84sw8t/K8l/7+7r3X0zyWeS/M35xgKA6ZllGcg/TfLuqnpdkv+b5H1JJrPGY1UN2U93D9kPAAfXrqXb3V+qqk8neSzJC0m+kuSReQcbZa9lWFUKFIDbMtMFD7r7QpILc84CAJNmRSoAGETpAsAgShcABlG6ADCI0gWAQZQuAAyidAFgEKULAIMoXQAYROkCwCBKFwAGUboAMIjSBYBBlC4ADKJ0AWAQpQsAgyhdABhE6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYJBdS7eq3l5Vj7/s9lxVfWxEOACYkrt226C7/zjJO5KkqpaSfCPJZ+ecCwAmZ6/Dy+9L8t+6++l5hAGAKdtr6d6fZPP7/UFVna2qraraun79+p0nA4CJmbl0q+o1ST6U5D9+vz/v7ke6e62711ZWVvYrHwBMxl6OdH8myWPd/T/nFQYApmzXiVQv80B+wNAywF5V1ZD9dPeQ/cAsZirdqnp9kvcn+QfzjQMcFbdThlWlRDnUZird7v7zJH91zlkAYNKsSAUAgyhdABhE6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYBClCwCDKF0AGETpAsAgShcABlG6ADDIXq6ne+CdOHEi29vbc9/PvK8Dury8nBs3bsx1HwCMN6nS3d7ensS1Nkdd3BuAsQwvA8AgShcABlG6ADCI0gWAQZQuAAyidAFgEKULAIPMVLpV9aaq+nRVfbWqrlbV35h3MACYmlkXx/hXSX6ru/9uVb0myevmmAkAJmnX0q2qNyb56SR/P0m6+7tJvjvfWAAwPbMML78tyfUk/7aqvlJVn6iq1885FwBMziyle1eSn0zyb7r7nUn+PMkvvXqjqjpbVVtVtXX9+vV9jgkAh98spXstybXu/tLO95/OrRJ+he5+pLvXunttZWVlPzMCwCTsWrrd/T+S/FlVvX3nrvcleWquqQBggmadvXwuyaM7M5e/luTn5xcJAKZpptLt7seTrM05CwBMmhWpAGCQWYeXD4W+cHfy4BsXHeOO9YW7Fx0BgDmYVOnWQ8+luxcd445VVfrBRacAYL8ZXgaAQZQuAAyidAFgEKULAIMoXQAYROkCwCBKFwAGUboAMIjSBYBBlC4ADKJ0AWAQpQsAgyhdABhE6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYBClCwCD3DXLRlX19STfTvJikhe6e22eoe5EVS06wh1bXl5edATYkxMnTmR7e3vIvub9O768vJwbN27MdR8cXTOV7o4z3f2tuSXZB909931U1ZD9wGGyvb09md+LKfzDnYPL8DIADDJr6XaS366qK1V19vttUFVnq2qrqrauX7++fwkBYCJmLd37uvsnk/xMkl+oqp9+9Qbd/Uh3r3X32srKyr6GBIApmKl0u/sbO/99Nslnk7xrnqEAYIp2Ld2qen1VveF7Xyf520memHcwAJiaWWYv/7Ukn92Z0XdXkn/f3b8111QAMEG7lm53fy3JTwzIAgCT5pQhABhE6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYBClCwCDKF0AGETpAsAgSpcjYXNzM6dPn87S0lJOnz6dzc3NRUcCjqBZLngAh9rm5mbOnz+fjY2N3Hfffbl8+XLW19eTJA888MCC0wFHiSNdJu/ixYvZ2NjImTNncuzYsZw5cyYbGxu5ePHioqMBR0x1974/6draWm9tbe378x4EVZV5/MyYn6WlpTz//PM5duzYS/fdvHkzx48fz4svvrjAZNMxpd+LKb0WxqmqK929ttt2jnSZvNXV1Vy+fPkV912+fDmrq6sLSgQcVT7TZfLOnz+f9fX1/+8zXcPL+6cv3J08+MZFx9gXfeHuRUdgwo586VbVkMcYrlqc702WOnfuXK5evZrV1dVcvHjRJKp9VA89N5n/x6sq/eCiUzBVR750p/IXBQAH35EvXabPKUPAQWEiFZPnlCHgoFC6TN7Vq1dz7dq1V6xIde3atVy9enXR0YAjxvAyk3fPPffk4x//eB599NGXhpc//OEP55577ll0NOCIcaTLkfDqCXMm0AGLMHPpVtVSVX2lqj43z0Cw35555pk8/PDDOXfuXI4fP55z587l4YcfzjPPPLPoaMARs5fh5Y8muZrEmeMcKqurq7n33nvzxBNPvHTfpUuXrEgFDDfTkW5V3ZvkZ5N8Yr5xYP99b0WqS5cu5ebNm7l06VLW19dz/vz5RUcDjpiZLnhQVZ9O8i+SvCHJP+nuD36fbc4mOZskJ0+e/Kmnn356n6PCK93OymC3y2fAP9yULhIwpdfCOPt2wYOq+mCSZ7v7yg/brrsf6e617l5bWVnZQ1S4Pd2959udPA7gTs0yvPyeJB+qqq8n+VSS91bVr881FQBM0K6l292/3N33dvepJPcn+Z3u/sjckwHAxDhPFwAG2dOKVN39u0l+dy5JAGDiHOkCwCBKFwAGUboAMIjSBYBBlC4ADKJ0AWAQpQsAgyhdABhE6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYBClCwCDKF0AGGRPF7GHeTpx4kS2t7fnvp+qmuvzLy8v58aNG3PdB3A4KV0OjO3t7XT3omPcsXmXOnB4GV4GgEGULgAMonQBYBClCwCD7Fq6VXW8qn6/qv6gqp6sqodGBAOAqZll9vJfJHlvd3+nqo4luVxV/6m7/+ucswHApOxaun3rHI7v7Hx7bOd2+M/rAIDBZvpMt6qWqurxJM8m+UJ3f2m+sQBgemZaHKO7X0zyjqp6U5LPVtXp7n7i5dtU1dkkZ5Pk5MmT+x6U6esLdycPvnHRMe5YX7h70RGAA6r2ugJQVf3zJP+nu3/lB22ztrbWW1tbd5qNI6aqJrMi1RRex15M6TVP6bUwTlVd6e613babZfbyys4RbqrqtUnen+Srdx4RAI6WWYaXfzTJv6uqpdwq6f/Q3Z+bbywAmJ5ZZi//YZJ3DsgCAJNmRSoAGMSl/ThQpnBZvOXl5UVHAA4opcuBMWLGqJmpwCIZXgaAQZQuAAyidAFgEKULAIMoXQAYROkCwCBKFwAGcZ4uAHs2aiGbqZ1Xr3QB2LPbuCzs5Ar0dhheBoBBlC4ADKJ0AWAQpQsAgyhdABhE6QLAIEoXAAZRugAwiMUxOLRud0Wc23mck/qB/aB0ObQUIXDYGF4GgEF2Ld2qemtVXaqqp6rqyar66IhgsJ82Nzdz+vTpLC0t5fTp09nc3Fx0JOAImmV4+YUkv9jdj1XVG5JcqaovdPdTc84G+2JzczPnz5/PxsZG7rvvvly+fDnr6+tJkgceeGDB6YCjZNcj3e7+Znc/tvP1t5NcTfKWeQeD/XLx4sVsbGzkzJkzOXbsWM6cOZONjY1cvHhx0dGAI6b2Mhmlqk4l+WKS09393Kv+7GySs0ly8uTJn3r66af3LyXcgaWlpTz//PM5duzYS/fdvHkzx48fz4svvrjAZNMxpcu2Tem1HCRT/7lW1ZXuXtttu5knUlXVjyT5jSQfe3XhJkl3P9Lda929trKysre0MEerq6u5fPnyK+67fPlyVldXF5QIOKpmKt2qOpZbhftod39mvpFgf50/fz7r6+u5dOlSbt68mUuXLmV9fT3nz59fdDTgiNl1IlXdWklgI8nV7v7V+UeC/fW9yVLnzp3L1atXs7q6mosXL5pEBQy362e6VXVfkv+S5I+S/OXO3f+suz//gx6ztrbWW1tb+xYSONim9HndlF7LQTL1n+usn+nueqTb3ZeT3N56ewDAS6xIBQCDWHsZ4Ig7ceJEtre3576f271IyayWl5dz48aNue7jTildgCNue3t7Ep+3zrvU94PhZQAYROkCwCBKFwAGUboAMIiJVMC+OAyTWGaxvLy86AhMmNIF7tioma9TX9WI6TO8DACDKF0AGETpAsAgShcABlG6ADCI0gWAQZQuAAziPF2AI64v3J08+MZFx7hjfeHuRUfYldIFOOLqoecmsehIVaUfXHSKH87wMgAMonQBYBClCwCDKF0AGETpAsAgu5ZuVX2yqp6tqidGBAKAqZrlSPfXknxgzjkAYPJ2Ld3u/mKSGwOyAMCk7dviGFV1NsnZJDl58uR+PS0wUVU15HFTWPSB6di3iVTd/Uh3r3X32srKyn49LTBR3T3kBgeJ2csAMIjSBYBBZjllaDPJ7yV5e1Vdq6r1+ccCgOnZdSJVdz8wIggATJ3hZQAYROkCwCBKFwAGUboAMIjSBYBBlC4ADKJ0AWAQpQsAgyhdABhE6QLAIEoXAAZRugAwiNIFgEGULgAMonQBYBClCwCD7HoRewCmr6oWHeGOLS8vLzrCrpQuwBHX3XPfR1UN2c9BZ3gZAAZRugAwiNIFgEGULgAMMlPpVtUHquqPq+pPquqX5h0KgIOtqvZ0u53HTGFG9avtOnu5qpaS/Osk709yLcmXq+o3u/upeYcD4GAyE/n2zHKk+64kf9LdX+vu7yb5VJKfm28sAJieWUr3LUn+7GXfX9u5DwDYg32bSFVVZ6tqq6q2rl+/vl9PCwCTMUvpfiPJW1/2/b07971Cdz/S3WvdvbaysrJf+QBgMmYp3S8n+bGqeltVvSbJ/Ul+c76xAGB6dp293N0vVNU/TPKfkywl+WR3Pzn3ZAAwMTNd8KC7P5/k83POAgCTZkUqABhE6QLAIEoXAAZRugAwSM1j/cyqup7k6X1/4oPhzUm+tegQ3Dbv3+Hm/Tu8pv7e/fXu3nWRirmU7pRV1VZ3ry06B7fH+3e4ef8OL+/dLYaXAWAQpQsAgyjdvXtk0QG4I96/w837d3h57+IzXQAYxpEuAAyidAFgEKU7o6r6ZFU9W1VPLDoLe1dVb62qS1X1VFU9WVUfXXQmZlNVx6vq96vqD3beu4cWnYm9qaqlqvpKVX1u0VkWTenO7teSfGDRIbhtLyT5xe7+8STvTvILVfXjC87EbP4iyXu7+yeSvCPJB6rq3QvOxN58NMnVRYc4CJTujLr7i0luLDoHt6e7v9ndj+18/e3c+gvgLYtNxSz6lu/sfHts52YG6CFRVfcm+dkkn1h0loNA6XLkVNWpJO9M8qXFJmFWO8OTjyd5NskXutt7d3j8yyT/NMlfLjrIQaB0OVKq6keS/EaSj3X3c4vOw2y6+8XufkeSe5O8q6pOLzoTu6uqDyZ5truvLDrLQaF0OTKq6lhuFe6j3f2ZRedh77r7fyW5FPMrDov3JPlQVX09yaeSvLeqfn2xkRZL6XIkVFUl2Uhytbt/ddF5mF1VrVTVm3a+fm2S9yf56mJTMYvu/uXuvre7TyW5P8nvdPdHFhxroZTujKpqM8nvJXl7VV2rqvVFZ2JP3pPk7+XWv7Qf37n9nUWHYiY/muRSVf1hki/n1me6R/7UEw4ny0ACwCCOdAFgEKULAIMoXQAYROkCwCBKFwAGUboAMIjSBYBB/h/d89BmTQ1aMgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a figure instance\n",
    "fig = plt.figure(figsize=(8, 5))\n",
    "# Create an axes instance\n",
    "ax = fig.add_subplot(111)\n",
    "data = []\n",
    "for j in range(X.shape[1]):\n",
    "    data.append(X[:,j])\n",
    "bp = ax.boxplot(data)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Entrenamiento de Modelos"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**kFold - Linear Discriminant Analysis**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "kf = KFold(n_splits=10)\n",
    "resultados_train = []\n",
    "resultados_test = []\n",
    "for tipo in ('lsqr', 'eigen', 'svd'):\n",
    "    for _ in range(3):\n",
    "        for train_index, test_index in kf.split(X):\n",
    "            X_train, X_test = X[train_index], X[test_index]\n",
    "            y_train, y_test = y[train_index], y[test_index]\n",
    "            alg = LinearDiscriminantAnalysis(solver=tipo, shrinkage=None)\n",
    "            alg.fit(X_train, y_train)\n",
    "            resultados_train.append(alg.score(X_train, y_train))\n",
    "            resultados_test.append(alg.score(X_test, y_test))\n",
    "resultados_train = np.reshape(resultados_train,(30,3),order='F')\n",
    "resultados_test = np.reshape(resultados_test,(30,3),order='F')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Test de normalidad**  \n",
    "Rechazamos normalidad"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6.841526012379973e-09\n",
      "1.5443413303728448e-07\n",
      "6.841526012379973e-09\n"
     ]
    }
   ],
   "source": [
    "print(stats.shapiro(resultados_test[:,0])[1])\n",
    "print(stats.shapiro(resultados_test[:,1])[1])\n",
    "print(stats.shapiro(resultados_test[:,2])[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Test de Kruskal-Wallis**  \n",
    "No se puede rechazar que las precisiones sean similares"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KruskalResult(statistic=0.8802197802197335, pvalue=0.6439656517385672)\n"
     ]
    }
   ],
   "source": [
    "print(stats.kruskal(resultados_test[:,0],\n",
    "                    resultados_test[:,1],\n",
    "                    resultados_test[:,2]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**kFold - Quadratic Discriminant Analysis**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "kf = KFold(n_splits=10)\n",
    "resultados_trainq = []\n",
    "resultados_testq = []\n",
    "for _ in range(3):\n",
    "    for train_index, test_index in kf.split(X):\n",
    "        X_train, X_test = X[train_index], X[test_index]\n",
    "        y_train, y_test = y[train_index], y[test_index]\n",
    "        alg = QuadraticDiscriminantAnalysis(priors=np.array([1/3, 1/3, 1/3]))\n",
    "        alg.fit(X_train, y_train)\n",
    "        resultados_trainq.append(alg.score(X_train, y_train))\n",
    "        resultados_testq.append(alg.score(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Test de normalidad**  \n",
    "Rechazamos normalidad"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.5443413303728448e-07\n"
     ]
    }
   ],
   "source": [
    "print(stats.shapiro(resultados_testq)[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Test de Kruskal-Wallis**  \n",
    "No se puede rechazar que las precisiones sean similares"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KruskalResult(statistic=1.260000000000016, pvalue=0.7386526627497971)\n"
     ]
    }
   ],
   "source": [
    "print(stats.kruskal(resultados_test[:,0],\n",
    "                    resultados_test[:,1],\n",
    "                    resultados_test[:,2],\n",
    "                    resultados_testq))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Nos quedamos con el más sencillo**  \n",
    "El cual asumimos que es el Lineal Discriminant Analysis con lsqr (mínimos cuadrados)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "def performance_indexes(CM, PositiveClass):\n",
    "    classNum = len(CM)\n",
    "    TP = CM[PositiveClass, PositiveClass]\n",
    "    TN = 0\n",
    "    FP = 0\n",
    "    FN = 0\n",
    "    for real in range(classNum):\n",
    "        for predicted in range(classNum):\n",
    "            if (real != PositiveClass and predicted != PositiveClass):\n",
    "                TN += CM[real, predicted]\n",
    "            if (real != PositiveClass and predicted == PositiveClass):\n",
    "                FP += CM[real, predicted]\n",
    "            if (real == PositiveClass and predicted != PositiveClass):\n",
    "                FN += CM[real, predicted]\n",
    "    Sens = TP/(TP+FN)\n",
    "    Spec = TN/(TN+FP)\n",
    "    PPV = TP/(TP+FP)\n",
    "    NPV = TN/(TN+FN)\n",
    "    ACC = (TP+TN)/(TP+TN+FP+FN)\n",
    "    return Sens,Spec,PPV,NPV,ACC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sens: 1.00, Spec: 1.00, PPV: 1.00, NPV: 1.00, ACC: 1.00\n"
     ]
    }
   ],
   "source": [
    "X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.25)\n",
    "alg = LinearDiscriminantAnalysis(solver='lsqr', shrinkage=None)\n",
    "alg.fit(X_train, y_train)\n",
    "y_pred = alg.predict(X_test)\n",
    "CM = confusion_matrix(y_test, y_pred)\n",
    "Sens,Spec,PPV,NPV,ACC = performance_indexes(CM,1)\n",
    "print(('Sens: {:.2f}, Spec: {:.2f}, ' +\n",
    "      'PPV: {:.2f}, NPV: {:.2f}, ACC: {:.2f}').format(Sens,Spec,PPV,NPV,ACC))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**----------------------------------------------------------------------------------------------------------------------------**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Persistencia"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open('RForest_Zika_oversampled.bin','wb') as file:\n",
    "#     pickle.dump(alg_o, file)\n",
    "# with open('RForest_Zika_undersampled.bin','wb') as file:\n",
    "#     pickle.dump(alg_u, file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## FIN"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}