{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Задача: классификация стекла\n", "Часто на месте преступления остаются осколки разных видов стекол, которые можно использовать как улики, если определить тип стекла и от каких оно объектов. [Выборка](https://archive.ics.uci.edu/ml/machine-learning-databases/glass/) состоит из 9 признаков - химических параметров образцов и 214 объектов. Необходимо каждому образцу сопоставить один из 6 классов (например: стекло автомобиля, осколок посуды, окно здания) и сравнить качество работы решающего дерева и алгоритма решающего дерева и алгоритма k-ближайших соседей. В качестве функции ошибки использовать долю неправильных ответов классификатора. Дает ли масштабирование признаков значительное улучшение в качестве классификации? " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Домашнее задание 2" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Первое знакомство с данными" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "import sklearn as sc\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "from sklearn import tree, model_selection, metrics\n", "from sklearn.neighbors import KNeighborsClassifier\n", "from sklearn.preprocessing import StandardScaler\n", "plt.rcParams['font.family'] = 'serif'\n", "plt.rcParams['font.serif'] = 'FreeSerif'\n", "plt.rcParams['lines.linewidth'] = 2\n", "plt.rcParams['lines.markersize'] = 12\n", "plt.rcParams['xtick.labelsize'] = 24\n", "plt.rcParams['ytick.labelsize'] = 24\n", "plt.rcParams['legend.fontsize'] = 24\n", "plt.rcParams['axes.titlesize'] = 26\n", "plt.rcParams['axes.labelsize'] = 24" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "columnNames = ['Id','RI', 'Na', 'Mg', 'Al', 'Si', 'K', 'Ca', 'Ba', 'Fe', 'Type of glass'] \n", "data = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/glass/glass.data', \n", " names = columnNames, header=None)" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
RINaMgAlSiKCaBaFeType of glass
01.5210113.644.491.1071.780.068.750.00.01
11.5176113.893.601.3672.730.487.830.00.01
21.5161813.533.551.5472.990.397.780.00.01
31.5176613.213.691.2972.610.578.220.00.01
41.5174213.273.621.2473.080.558.070.00.01
\n", "
" ], "text/plain": [ " RI Na Mg Al Si K Ca Ba Fe Type of glass\n", "0 1.52101 13.64 4.49 1.10 71.78 0.06 8.75 0.0 0.0 1\n", "1 1.51761 13.89 3.60 1.36 72.73 0.48 7.83 0.0 0.0 1\n", "2 1.51618 13.53 3.55 1.54 72.99 0.39 7.78 0.0 0.0 1\n", "3 1.51766 13.21 3.69 1.29 72.61 0.57 8.22 0.0 0.0 1\n", "4 1.51742 13.27 3.62 1.24 73.08 0.55 8.07 0.0 0.0 1" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df = data.drop(list(data)[0], 1)\n", "df.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Посмотрим на соотношение классов." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "1 - 70\n", "2 - 76\n", "3 - 17\n", "4 - 0\n", "5 - 13\n", "6 - 9\n", "7 - 29\n" ] } ], "source": [ "X = df.drop(list(df)[9], 1)\n", "target = df['Type of glass']\n", "for i in range(1, 8):\n", " print(\"{0} - {1}\".format(i, sum([target[j] == i for j in range(target.shape[0])])))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Обучение моделей\n", "В качестве критерия сравнения моделей решающего дерева и метода ближайших соседей будем сравнивать доли неправильных ошибок, усреднённых по кросс-валидации с 3 фолдами. Для воспроизводимости результата зафиксируем seed." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.4181989705565946" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf1 = tree.DecisionTreeClassifier()\n", "np.random.seed(0)\n", "c_v_s = model_selection.cross_val_score(clf1, X, target, cv = 3)\n", "1 - c_v_s.mean()" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.36784947838224025" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "res = []\n", "k = [i for i in range(1, 20)]\n", "for i in range(1, 20):\n", " clf2 = KNeighborsClassifier(n_neighbors=i)\n", " c_v_s = model_selection.cross_val_score(clf2, X, target, cv = 3)\n", " res.append(1 - c_v_s.mean())\n", "min(res)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Масштабирование признаков\n", "Сравним результаты работы моделей после масштабирования признаков. Для метода ближайших соседей построим график зависимости доли ошибок классификатора от числа соседей." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.42270347506109907" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "np.random.seed(0)\n", "scaler = StandardScaler()\n", "scaler.fit(X, target)\n", "X_scaled = scaler.transform(X)\n", "clf1 = tree.DecisionTreeClassifier()\n", "c_v_s = model_selection.cross_val_score(clf1, X_scaled, target, cv = 3)\n", "1 - c_v_s.mean()" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.36784947838224025" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "res1 = []\n", "for i in range(1, 20):\n", " clf2 = KNeighborsClassifier(n_neighbors=i)\n", " c_v_s = model_selection.cross_val_score(clf2, X, target, cv = 3)\n", " res1.append(1 - c_v_s.mean())\n", "min(res1)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(figsize=(12, 8))\n", "plt.xlabel('Number of neighbors, $k$')\n", "plt.ylabel('Error rate, $Q$')\n", "plt.plot(k, res, 'g-', marker='o')\n", "plt.plot(k, res1, 'b-', marker='o')\n", "ax.set_title('Зависимость ошибки от числа соседей')\n", "plt.grid()\n", "plt.show()\n", "fig.savefig('1.svg')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Выводы\n", "Видим, что качество алгоритма метода ближайших соседей при кросс-валидации с количеством фолдов = 3 будет наилучшим при k = 14 и при таких условиях эффективнее алгоритма решающего дерева с параметрами по умолчанию. Как и следовало ожидать, масштабирование признаков не изменяет качество работы алгоритма ближайших соседей, а для решающего дерева даже немного ухудшает." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Домашнее задание 3" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Построим график зависимости средней ошибки и его стандартного отклонения алгоритма ближайших соседей от объема выборки на обучении и контроле. В качестве выборки возьмём первые $m$ элементов выборки. Для этого $K = 20$ раз случайно разобьём выборку на обучение и контроль в отношении $1:1$. Модель ближайших соседей имеет единственный параметр - число соседей, так что для каждого разбиения на обучение и контроль подберём нужное число соседей, минимизирующее долю неправильных ответов классификатора на обучении и контроле соответственно. Для этого выберем минимальное целое число в отрезке $[1, size(X_{train})]$, которое будет доставлять минимум функции ошибки на обучении или контроле. Получим два набора из 20 ошибок. Вычислим среднюю ошибку и стандартное отклонение. Повторим процедуру для $m \\in [2, 100]$. После этого построим график средней ошибки и стандартного отклонения от $m$." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "from sklearn.model_selection import train_test_split\n", "K = 20\n", "error_mean = []\n", "error_std = []\n", "for m in range(2, 101):\n", " X_data = X[0:m]\n", " y_data = target[0:m]\n", " error_data = []\n", " for k in range(K):\n", " X_data_train, X_data_test, y_train, y_test = train_test_split(X_data, y_data, test_size=0.5, \n", " random_state=k)\n", " error = []\n", " for n in range(1, X_data_train.shape[0] + 1):\n", " clf = KNeighborsClassifier(n_neighbors=n)\n", " clf.fit(X_data_train, y_train)\n", " predictions = clf.predict(X_data_test)\n", " error.append(1 - metrics.accuracy_score(y_test, predictions))\n", " error_data.append(min(error))\n", " error_mean.append(np.array(error_data).mean())\n", " error_std.append(np.array(error_data).std())" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwkAAAIRCAYAAAABafHOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3XmcHVWZ//HPk50kLCEJIWxp9iUBWQIIKERAZNRRVCCgqAExI4KII0qU0VFkkBkXxIFBNtH8REBxQYhEI9DshkVIDElYAkkggBIIIQtZ+/n9cU51163U3fp2973d/X2/XvdVt6tOVT23qu7teqrOqWPujoiIiIiISKJPvQMQEREREZHGoiRBREREREQKKEkQEREREZECShJERERERKSAkgQRERERESmgJEFERERERAooSRARkU5lwSfqHUdPpm3cWLQ/pCdQkiAiIp3CzHY1s/OBJ4GD6h1PT6Rt3Fi0P6QnMXWmJlI7M/s98OHUqNXAQmA5sDnwJnAHcIu7L+zq+ES6ipkNBc4FTgH2jaNfBw5098V1C6wH0TZuLNof0lPpToJIx1oKnAGMdPex7n44cDDwW+C7wFwzO62eAYp0FjM7Cnga+C/CydIi4GrgIJ0sdQxt48ai/SE9me4kiHSAeCfhcOBod59TpMxlwHnxz0Pc/dGuik+ks5nZe4A/Ai8A1wN/AJ5z/ZPpMNrGjUX7Q3q6fvUOQKQHuaRYghDdQ1uScCSgJEF6BDPrD/wf8O/ANe6+sc4h9Tjaxo1F+0N6AyUJIh1jNfCzMmW2S71/rPNCEelyk4Gr3P2qegfSg2kbNxbtD+n53F0vvfTq5BewLaHeqgPTgX6Z6XsD1xHuLjwQy94M7JAqsxVwKSEh8Ti8D3gceCbOez4wJDXP+cCSWN6BPwPvidNOBWanps0DTs/ENRS4JM73YFzPT4E94vSmGHdLXMYa4EfAcKA/cAXwVmodvwLeAXwIeCI1/iWgb5Ftd2Wq3BLgq5npW8bt8kfg/rjtfg28t8T+2Ab4X2AGMDPO8yNguzj9W3F9c4F7CY3QHXgVaCY8uWQj8HtCPeRfpmJ8A7gYGFTD8XJ83FZ3AM/Fz/VdYItMudz9C+xSYr98FHgkNc984DPt3Zcxjj8BAwl3yP5AOBZnAg8DlwOjMnGfDTyfWtYs4FNxWp84z+vAD1PznE44RpN5ZgIfjdPeCzyUOU7O74Tt7YSHEDwcj4NkfX8lfG//Ef/ev8L17kH4Pv0xHlNPEKqt7JZTttpt/CTwdtzXi1Ox3g5MqCC2iYS69X8kHP93pj8XsBNwQ2q5DhxXZFkHp8q0xONslzjNgJOBuwi/Z7Pi66sU+U3ILLsf8M14vDiwIe6LRwnH9izg28CI1DznAS+mYroTOCZn2RdS+P26EziqPfsjtUwDPg78DphG+B7cBXwNGFjkt+prQHP8+0NxXasIv/1fBfqktsWPgRWpmKcBh8fpwwn/f14FPp6a5xLCcZ3+bh8Qp3+Bwu/qTOCE9v626dW9XnUPQC+9evKLcGL/aWBB/If9P8BmmTLvBFYC1yT/FIH9gHWEk9Q+mfLXxB/rX6TGDQauiuN/myn/iTh+JTAgM22v1I//4ZlpBxKSgq+k4jqBthPhPVJlZ8Tx1+Rsg0vjtGeI7aDi+D2AZYSTUQc+kDPvIOCVWM6Tf2yp6YcRToCuAvrHcdsQTuydcJKT3X7vJZz0nEZbu6zzYvmFwAhCkvDV1DzfitN/lhr3QeD38b3Fz+fA12o4XvrGmF8BDk6NO5Vw8rMYeGdmntNK7N+/5O0XYIcS+72qfQmMJJxcn0pIIo5PTTuE8ISvpdn9C7w7FUM6GT6F8F1xYElmniNS82yXmTY4Ne3j2dg7YnsTThi3Tv2drK8pdbw+QAVJAvDZuL0mpbbX/jGW1cCZqbJVb2NCcrBnfH98jHMF8XtSIq5+hLuil9J28rlPPL7W5RwvD8WYHfh1kWX+H+HE1IE/Z6Z9L45/V+q7lBxnU6r47nw9zvNAalx/4D/i+EdJJR2U+F0s9/1qz/6I04YSLizMJyaCwABClSUH5lD4u3oDsJa236Zz4vvXUseeEy4opX9bvxHHP5sZ/xPavls3ZmK7kJzf6czxs8n3Tq+e/ap7AHrp1RNfhBPgV1I/rK/Hf1bb5JS9MZb598z4x+L47EnhRWROWOP4EbRdqds8Nf7oOH5xzrq/mIpxl9T4wYQrqldkyveh7arklNT4qXHcRZnyAwlXWR24LzOtf/zH97M4/Xc58U2M22dhLHN0atpQwpNE5pO54kh47Gwyz9mp8aMICc75mfJbEk4CnHCS+i0KTwa/ld3mhJOZK1J/3xfLnJ79HFUcN8lJwCdypn2BtpOFoRXu32L7pU/efm/PviQ02L85br+v58RwIm1X4LdLjd8lFYOl1vFg3B+nA+/OLCs9Tzb5+3Bq2tHZODpiewOXZ8oUJAlx3JmUSRIIycAG4Nqcae+g7Yr4/u3dxsD3UtMnxOmLKtgmnyd8r/pnxv8fbSfb6RPPG2n7fqwj8xtHSJxeItyVcmBqatqQOI9T+H0bF8fNr+K7c0acpznnWE+urO+T870puU3I+X7VcMx/P44/ImeeH9B2pT65KDOQ8NQkJ1xMuZaQsBmwPSEhTY7B9+Vsi3tT4/Yi3L3YgpCU7Fxk+92XE9tlFPne6dWzX3oEqkgncPdnCD/i2xPuJLwJfAdYaGZfzRSfTqhmcHdm/Mo43DEzvqXIareOwxWEfyjZ8gXzmdnhhJOkvOVeCOxK+OfQyt1bgE8S/ln9stw6gB8Co/Omufv6+PaKOPxXMxtNodMJV3rzYvwKocrDrzzTaNDdV6Tmu9TMNk/FM4RwwpMuv5ywn64k7I+p7v4GJbi7U7h9kti81HzFmNlWhNv+LYTqUlnXE65QjiFUM8quN++4yJ0W92O2TLnlFduXIwlXGodQeEwk67qVUJ1rS8Ln22S9cVtCSFp/7+7L3f0Gd7+/SGwFn8HMdidU8XirSOybaOf2vrzccoHbCE+7KeVywh2Lm7IT3H0WodpYX8KJNbRvG+fFWvLYjM/7vwiYkfp+JpJ9MZ7wvUusJ9y5e4qQ+H8qM98J8fO8Gf9O75vVhKo6t6amQ/HfvlKK7fPBhJNtJ1wgyJYv933N+z5UvT/iMXoe8KK7P5iznh/FdRxC+I3F3dcSqrVB+F3/nLvP9WAJ8H7ajvm83/L0Z/sB8F13f8vdf+ju2WO02P+Jk4CPpD5b2e+W9BxKEkQ6ibu3uPvL7j6VcGVsDrAZ8N/pRMHd/5+7H+juTwKY2XZmNgXYPRYZUmo9FuwD/Jzwz/VTOf/gs/OMAv4f4Xb5Jssj/KNfQ87Jjrvf6+6TvcwzwM3sE8ABhCthRbn7Y4Qr1H0JJ+rJ/DsQkqzsiWLiqDhcWGR6cuI3FNjPzIYQrvC94O6rc+L4jbuf4+5vuvvzpWJOzbOgknIV2p9wB2SJu6/LWddqQhUHCNVuukyZfTmYcDLkhHreeW6Lw6Jxx8dJHkpoK1JNbIMJ+/pcQrW0SlW9vSs5Ltz9tZh0Fot3AOFKNJQ/dt9pZv1oxzZ295fKxZpjX0K99WPNrDn9IiRwi+Jr38x8Ttt+OzP+hiROJ1Rz2UQ82T3R3U9y95b4W/YuwkUKCJ+73cxsJ9raC33e3V+tZXkp7TnmjyD8xi3MK+zuLxLaM6TngZCEAazKuRjyFuHOH4R2bZsws75m9iXgb+7+SJFYc5nZXoTqYJOqmU96DiUJIl3A3d8G/i016tvxSmYrM9vXzK4lVFe4kVCftJQTzexxQrWmpwgnw6cSrswVFU86biJcVcp7ytIWhDrrq9t71cjMxhEa755MqE5QTnI34UwzS36XPgVcl7rKnDUuDjcrMj2pFgWhjcduhPq/qyqIpxaXmdlTZvagmf3OzM40s2IxppX7PNCWtO1XW4iVq2BfJidJRti+eZ6Kw91ispZdxxLCnbSlhJPUSmMzQnuUO9y95HGfo17bew/anixYbN3JegcSLhbUvI2jHczsOTO718z+ZGY/NrOxqel7xeEf3X1C5vVOd2+Krztyln0j4c7LnsC7AMxsR8L+LPm4ZzPrZ2anEi5c7EioltNe7zSzJ81sIW0JzWRCW648O5rZYjObaWZ3mdkVZpZNgrLasz8663h7Mg63y5l2KKH9wn8By2NCXZF4V+nXhP9bC6uIR3oQJQkiXecx2q4KDSLeKYhXen5EOHH/trtfFK8qlfNbdz+IUAVkN+BvhCeX3GdmW5eY72LC1fTrikwfGIdbl1lOLjPbkvB0jDOquJp5K/BPQhWno+LJ30TarpLlSa6q5f1zJF4dfi3++Tptn2u3zJXOjvYfhBOC9xAadV4LPGFmw8rMl3yeEWY2sEiZl+Pw9ZqjrECF+3JR6v02Rcr8Mw5XERpOZu1A2G6TgQfNbNsKQ/w3wv7/zwrLp9Vre6evBuceu6n1JuvuiG0M4YLCPu5+FHAW4RidY2bJBYyRcbhbkfmLcveVhCc1QbjQAeUTfcysifDbNR6Y5O43ERrrttcjhLteOxPuRN5KaPc0x8zG5JR/hdBY+FDatslsMzurxDrasz9K/l5F7Tneknmey5k2k5B0/YVwR+DX8SJRSfH38TpCVc4/VRGL9DBKEkS6znoK2wosjMOfE27lf6HKKgIt0HrLfgHhtv5fCVfxLsqbwcw+DBxLYf3VrNcIVwQB3lesUKyytMlowonCDe5+T7kPkIh1b5M2BJ8l3G5/wt1LVR+ZHYe5V/3iP7ot45+zaLszswXhiVK5inyuaqyK+2Qd4YroHMLV1S+XmW926v3YImWSu0+zaguxIpXuy5dpS34PL1JmaBz+Pe/uVDyJ/C7hpGYMcGO5RM7MDiZs01Oz1TAqVK/t/SxtJ8HFrlgn633V3f9JB2zjaGNStSpWnUqSq/+NV5mT6lTjYmdhueLdpTxJW5+TYlL8McIdhmLL2ZHwm7U58BV331CsbBVavM3L7v4VQqK7N6HNUdYGd18DrW3JvhnH/7jElff27I/keBttZiOLzNOe4y258zY3b6K7ryIk0ysIbRi+XsEyv0CIv2RVUen5lCSI1MjM3mVmR1dQdBfCP0OAhe7+mpntTHgUH4QrWmnJP6iKvqfxH1FSp3V8TpFhhIa2J8XqT8WW44SG1AAXFakeciht9YbTJsVhe6oLXE242vYxwsnf1aWL85c4PL7IHY/tCXcP5gHPxYRjYZz233knQTGJ+kz1oeeL23Je/HOvUmUJJwbJFcRN2opEu8ZhtVVr2mNSHJbcl/EE/Y/xz2JJZXISXjTuePwmTxs6mvDoyVJuJDzqdGmZcsXUZXvHE+Ek6apovR21jXPMj8P+hN+n5ERzNOFEcRNmdgRwUt40d3+O0JfAZoSE4VEPDxEo5jOEJ479I5PYtJ6cp6of1iJpKJz3u5iVfF/7EbbJJtq5Px4gVNdL+oXIk+z32yuIM5G0RdikEXzC3V+hLSH8emxEXcyehCdcfVKNlEVJgkjtHgW+bGYfKlPuW6n3yQ92ugrK58xsgJmNMrMvEuouA+xsZkNSJ8KlbhcnP/7pq6TJ93wLwuNAWxsjZ/4Bp98ndyJ2Ax42s8PMrH9sWHg8oQrNd3LmXUt4BKjnTCv4vYnrbh3n4WkdvyHU8d2VcIWx1DJ+RHimd3/CnZisSXF4VupK87fj8N3ADDPbL1b36hsb536NzBOdov6ZYZ5N9ku8Gp6caJSslx2rayR3Gz6drXIT6wifSDjBvLnUequUnb/qfUl4gosDH8/W547VG04iPO3lh3nrTapAxDYySd3xH+Qkf+n1/sDds9u0WHybqGF7J9PTx0Kp4yLPeYQTxvFFLjBMIiQw6au+7dnGiWLHSHJsLgeed/e5hJN8gIvN7DwLDa0xs4Fm9lnCs/vT35GC7zFtDZhPoTDRz9s3yf49MF5sMTM7kPA9TOwc7ziUU+3vYrlt8iahf5tk32bLV7U/3H0RbU+eOid78SVWhXoP4c7dQzlxDcreXYv7ZhLh6VLp5LBfZgih3ddThAsn/5dzpy7ZL1sBp2Tu4rbusw5K2qS78AZ4DqteenX3F+GH98eE+q8TSD1LmpAI/IS254ifl5q2GeHWdfIM6jdjmc/T9nxsJ/Tkunec54Y47jep5fQl1ANOHvN3QGrax+P4tWz67PPtU+s4LDPt+6lpTqhXu4rQKdLBmbJ/jmV+mrNtkk56nqbw+erbkenIiLYOts7KbNtVcfypmWXvT6iL2wJMTI0/jtAI9uxM+T7ALZnPtTJum2eAXYvs36TvgHuLTDdCo0MnVBtLxn82jnuMTCd6JZaTbPeHgK3i+MGETpgeAnbMzPPpWP6VnOXdHKd9NzN+UOrzH1nrvozTkj43nqDtWf2bEU5O5gHjMuU/lorh4NT44bT1nvuHzPFxeGqenXK+g7nHSUdu79S86T4b3tOO34xTCVVAlhKf3x+Pz68R+hU4MmeeqrZxzjHyBm0dpI0kNHpdD5yWKrs3hb0MryPUwV9D6ncoc7ycmvq7D6FK1cxMuWvj8v6UGvchCr+Lywjf531SMawE7qpge/5nLP94Zv9+kPD9XkNhXwLltknSI3HSf8DbZHpRr3Z/xGP0F3Ge3xB/EwjH/MOE431YZp4Jqe3zXWBw6jt8C6HX5ex34X9j+X9S2KdKelkXZOZJOqN7nk2/2+lODEdXe6zr1X1fdQ9AL7160ovQWO5/CI9OnBb/efydcPv4YjIdV8V5xhMa2y2P5SbE8fsSqiDNI9zS3pJw9f7t1A/2g4RHhM6J6/k+hT12fplwwpGUvxd4b5z2Edo6x3JCVYNJmdhOBP4U/5Eujf/gRqemNxGu/Lak/pFeG//p9SckOstS67iZ0FnUh+M/RSf0S5B8ZiOckA+Mf0+I60/mf5FQdzkd41DCFboH4ue5g9BQcb8i+8iAzxGeP/5W3MZXAFvklB1OuGOxNhXD9cDuqTJjCU9lSaavIlSFujN+tnOBflUeR8cQnizyBKEH5D/EfdkvU+4c2p7i5PE4ao6v+2jrzXp53PfJtNmpeeYQqva0a19m4plAqPbwRFzPnYS7UkMy5a4jnIgly1pNYedfn01Ne4lQJW8S4RhNxv+NcMWz2HHy5Y7e3qnyHyHcGUrW9wwwuR2/F7vEbfFY3F53EK48Dy8xT0XbuMQxMpfw29RM+K7tkzPPyLjvnyIcz3OA/6aww7P08VLwvSTUgT869R26jLZjsSXOtzPhu/iVuI8XE76Hw+N8FxKSqNvJnDhnYu1LSKySxNLj9ryPUKXsaULVp4Mq2Cb3EH47kgsyX6UwYbqD2DN0e/dHnOckwiNSnyT8RvyOUPXKcspOiOteRPgOPEz4rZtL+F+zWarsgLjMdOL1TwovoNyYmjaP8JCFiyn8bt+ebK8Y17zUtIeBD1d7rOvVPV9JL5ciIiIi0kDMbAIheVnk7k31jUZ6G9UtExERERGRAkoSRERERESkgJIEERERkcbUNw6L9ews0mmUJIiIiIg0mPiY0gPjn9vEfnVEukytz9eWDjBixAhvamqqeTmrVq1iyJBN+r2SHkr7u/fRPu9dtL97l/T+3rhxIwMGDGDdunXJ5L5m9vyQIUPYZZddGDhwYN3ilI5Rz+/3448/vtTdi/X83UpJQgNoamriscceq3k5zc3NTJgwofaApFvQ/u59tM97F+3v3kX7u3ep5/42s0WVlFN1IxERERERKaAkQURERERECihJEBERERGRAkoSRERERESkgJIEEREREREpoCRBREREREQKKEkQEREREZECShJERERERKSAkgQRERERESmgJEFERERERAooSRARERERkQJKEkREREREpICSBBERERERKaAkQURERERECjR0kmBm48zsFjObb2bzzOweMzumivn7mtk5ZvaEma0ysxfM7EYzayoxz05mdp2ZPWNmfzezmWZ2UonyW5vZZbH8k2Y228w+Z2ZW3acVEREREWkMDZskmNmxwCPAQmA/YCzwK2CGmZ1Zwfx9gJuAHwMjgb5AE/BxYK6ZHZgzz37A34BBwAFxvRcDt5jZRTnltwceB94BHOru+wOfAX4I/EyJgoiIiIh0Rw2ZJJjZloSEYC5wobuvc/cW4CfAbcBVZrZrmcVMBEYDO7v7DsBQ4CTgDWAzYGpmnX2BW4G3gbPcfZUHtwOXA98wsyMz6/g5sA3wSXdfBuDujwJfAz4FfLJdG0BEREREpI4aMkkAJgHDgJvdfUMy0t2dcHegH3BOsZnjFfzJwEfcfVGcd4O73wp8LhYbG+8EJD4I7A7c5u4rMou8KQ6/lFrH/sAxwD3uviRT/uakvO4miIiIiEh303BJQjypThKAu3KK3B2HZ5jZkCKLGUU42V+aM+02YH18Pzo1/gsl1vk48CbwITMbE8cVjdHd/wH8HdgfOKJIjCIiIiLSyzRNmcak6avqHUZZDZckEKrv7BbfP52dGE/8Xwe2APbMW4C7v+ruPyoybR2wCtgIPAeticnhJda5EVhA2F4HxdFHFCsfPROHhxSZLiIiIiLSkBoxSUiu1K8ntA/IsyxTtmJmti2wFTDN3d+Mo0cS2ilAuGNQcp0xqRhTaflqYxQRERERqad+9Q4gx05xuCy2QchTywn4KcAG4Bs560wvu9Q6h9OWVFRSfhNmNpnQboJRo0bR3NxcMuhKrFy5skOWI92D9nfvo33eu2h/9y7a371Po+/vRkwS1sVhS4kyybSqGgXHNgxfAv7D3WfnrLPUetPrXJ8zvqoY3f0a4BqA8ePH+4QJE4oHXqHm5mY6YjnSPWh/9z7a572L9nfvov3di0yfBtDw+7sRqxstjsOtSpQZFoeLqlz2RcCfgf8pss5S621dp7svB5ZXWr7KGEVERERE6qqRk4RBZjaoSJlhmbJlmdkpwI6EPhCy1ZiWAyvj+3In/Yszw0rLi4iIiIh0C42YJCwD5sX3TdmJsaO14cAKYH4lCzSzCcCHgE+k+11IxKTh/vjnzjnz94njnfA4VIAHipWPdonDRyqJUURERESkUTRckhBP2K+Ifx6bU2QCIe4b3H1lzvQCZnYYoYHwJHdfn5nWP9Vzc6l17guMAG539xfiuCuLlTezrYEDgdm0JR8iIiIiIt1CwyUJ0VTgDWBivIqfNpHwdKIrIPRxYGaXm9lUMxuYLmhmBxI6STs99o+QnjYgLiOZZzrhzsQJOdWcJsbhZckId3+K0L7hWDMbkSl/IqHB8g9LPKFJRERERKQhNWSSEO8QnAIcDFxoZv3MrI+ZnR7Hn+Xuz8biewDnAp8EjkqWYWbjCCfxhwGzzGx+6vU0oR3C3u4+N66zhZAMbA1cbmaDYgJyHPBV4Dvu3pwJ9Yy4nJ+b2eZxvfsDPyAkOlM7dsuIiIiIiHS+RnwEKgDuPsPMDgW+Dcwl3D14BTjW3e9OFV0A3A1sCcwEMLMdgL8Q2i4ML7GaqzPrnG1mBxGegjQHWB1fp7j7rTkxLonlvw08YWYrCInX+cA1uosgIiIiIt1RwyYJAO4+CzihTJkNwDGZcS8B27ZznS8Q7kpUWn4pcHZ71iUiIiIi0ogasrqRiIiIiIjUj5IEEREREREpoCRBREREREQKKEkQEREREZECShJERERERKSAkgQRERERESmgJEFERERERAooSRARERERkQJKEkREREREpICSBBERERERKaAkQURERERECihJEBERERGRAkoSRERERESkgJIEEREREREpoCRBREREREQKKEkQEREREZECShJERERERDpY05RpNE2ZVu8w2k1JgoiIiIiIFFCSICIiIiIiBZQkiIiIiIhIASUJIiIiIiJSQEmCiIiIiIgUUJIgIiIiIiIFlCSIiIiIiEgBJQkiIiIiIlJASYKIiIiIiBRQkiAiIiIiUkZ370G5WkoSRERERESkgJIEEREREREpoCRBREREREQKKEkQEREREZECShI6kZkNNbMj6x2HiIiIiEg1Gj5JMLNxZnaLmc03s3lmdo+ZHdPOZR1qZtOKnbib2UFm5mVes3Pm62dmC7JlgRVAS3tiFRERERGpl4ZOEszsWOARYCGwHzAW+BUww8zOrGI5h5nZdOCvwPsp/rknE07qFwJPZ17PxjJ/yJnvRGAksCTzugN4sNI4RUREREQaQb96B1CMmW1JSAjmAhe6+4Y4/ifAccBVZnaPuy8os5ydgMGEBOA5oH+RcpsDxwB7u/szOdPfDdwH/CYz3oALgPe6+8yqPqSIiIiISANq5DsJk4BhwM1JggDg7g7cREhwzim3EHdf7O53ufti4JUSRccBn85LEKKTgeeBJzPjjwPeUIIgIiIiIj1FQyYJ8ep8kgDclVPk7jg8w8yGVLFoLzrB/WF3z60aZGZ9CVWKbo1JStoFwNGxvcSVZnZ8jF9EREREpFtqyCQB2AbYLb5/OjvR3ZcCrwNbAHt2QTzvArZl06pGuwFJI+i9gM8DdwL3m9mhXRCXiIiIiEiHa9QkYUwcrgfeLlJmWaZsZzoZeBF4ND3S3Z8DBsYYJgMz4qQjgPvMbEIXxCYiIiIi0qEateHyTnG4LKd6T6JLkoRY1ehjwC/zYnH3jcBi4FrgWjM7nNDgenvgajPbK28+M5tMSCwYNWoUzc3NNce6cuXKDlmOdA/a372P9nnvov3du2h/dx/V7qdi5Rt9fzdqkrAuDkv1MZBM6+z6/+8GRpGpalSMuz9kZu8BZgN7AKOBl3PKXQNcAzB+/HifMGFCzYE2NzfTEcuR7kH7u/fRPu9dtL97F+3vbmD6NIDK91Ox8tUup04atbrR4jjcqkSZYXG4qJNjOZnwVKSHK53B3Z8l3FkAOKAzghIRERER6SyNniQMMrNBRcoMy5TtcGbWj1DV6LfuXm3PyQ/FYW6/DCIiIiIijapRk4RlwLz4vik7MXa0NhxYAczvxDiOJDxp6dZ2zLs+Dhd2WDQiIiIiIl2gIZOE2ND3ivjnsTlFJhBiv8HdV3ZiKCcDrwH3t2PedxAShFkdGZCIiIiISGdryCQhmgq8AUw0s2ycE4ENxETCgsvNbKqZDSyxzKTqT9nGzqmqRr+LTzDKK3OgmR2S7Twt3uk4C/hKiacziYiIiIjuEja9AAAgAElEQVQ0pIZNEuIdglOAg4ELzayfmfUxs9Pj+LNiA2EITxE6F/gkcFTe8sxsZ0KHaAC7VBDCBGAERaoaxUejPgjMBP5iZnvHZGVv4A/At9y9PdWURERERETqqmGTBAB3nwEcChwEzAXmAKcBx7r7damiC4C7gccJJ+2tzGyMmT0S508+73Vm9pKZnV1i9ScT2kY0F4ltI3AecB9wIKFK0gzgJOBEd7+y8k8qIiIiItI4GrWfhFbuPgs4oUyZDcAxRaYtAg5px3pbOzsrUeZq4Opqly0iIiIi0sga+k6CiIiIiIh0PSUJIiIiIiJSQEmCiIiIiIgUUJIgIiIiIiIFlCSIiIiIiEgBJQkiIiIiIlJASYKIiIiISDs1TZlG05Rp9Q6jwylJEBERERGRAkoSRERERESkgJIEEREREREpoCRBREREREQKKEkQEREREZECShJERERERKSAkgQRERERESmgJEFERERERAooSRARERERkQJKEkREREREpICSBBERERERKaAkQURERERECihJEBERERGRAkoSRERERESkgJIEEREREREpoCRBREREREQKKEkQEREREZECShJERERERKSAkgQRERERESmgJEFEREREJGqaMo2mKdPqHUbdKUkQEREREZECShJERERERKSAkgQRERERESmgJEFERERERAooSaiCmQ01syPrHYeIiIiISGfqV+8AyjGzccA3gHcADrwKXOzud7VjWYcC3wT+293vK1GuH/A0sEvO5HfnlN8pLvdIYC2wGvi+u/+62hhFREREpOc46Dszcsf/9m8v8dEDd+jiaCrX0HcSzOxY4BFgIbAfMBb4FTDDzM6sYjmHmdl04K/A+yn/uU8ERgJLMq87gAczy94P+BswCDggxnkxcIuZXVRpjCIiIiLSM7h76/vXV60reCXWbmipR2gVa9g7CWa2JSEhmAtc6O4b4vifAMcBV5nZPe6+oMxydgIGA5OB54D+ZcobcAHwXnefWaZsX+BW4G3gLHdfFSfdbmaXA98ws7+UumshIiIiIj3LayvWtr5/9MJjMWubNv7ivwDwkQO27+qwqtLIdxImAcOAm5MEAcBDanYTIcE5p9xC3H2xu9/l7ouBVypY73HAG+UShOiDwO7Abe6+IjPtpjj8UgXLEREREZEeYu4rb7W+H7n5QEYMbXslBvXvW4/QKtaQSUK8mp8kAHltD+6OwzPMbEgVi/byRbgAONrM5pnZlWZ2fIwnzxdKxPg48CbwITMbU0WMIiIiItKNzXsle+24+2nIJAHYBtgtvn86O9HdlwKvA1sAe3bUSs1sN0LjY4C9gM8DdwL3x0bP6bIGHF4ixo3AAsI2PqijYhQRERGRxjYvdSehu2rUJCG58r6eUN8/z7JM2Zq5+3PAwLjMyUDSHP0I4D4zm5AqPhLYLL5/s6tiFBEREZHG1hOShEZtuLxTHC7zdPPwQp1yAh7vACwGrgWuNbPDCQ2otweuNrO9Ykw7pWZbtumSSsdoZpMJiQijRo2iubm55thXrlzZIcuR7kH7u/fRPu9dtL97F+3vxlNsf5Qbv26js+C11e1eTqNo1CQheT5UqWdDJdOKtRfoEO7+kJm9B5gN7AGMBl6mLcZ0LFlFY3T3a4BrAMaPH+8TJkyoOdbm5mY6YjnSPWh/9z7a572L9nfvov3dtZqmTANg4aUf2HTi9DBtk/1R4fg5S5bTMuOB1skVL6fBNGp1o8VxuFWJMsPicFEnx4K7P0u4swChLwRoixGKx9llMYqIiIhI/c3tAVWNoPGThEFmNqhImWGZsp3toThM+llYDqyM78slCV0Vo4iIiIjUUU9ojwCNmyQsA+bF903ZibGjteHACmB+F8W0Pg4XQmt/DffHcTtnC5tZnzjeCY9DFREREZEeTklCJ4on4FfEP4/NKTKBEPsN7r4yZ3pneAchQZiVGlcqxn2BEcDt7v5C54YmIiIiIvXm7j2ijwRo0CQhmgq8AUyMV+XTJgIbiCfpFlxuZlPNbCDFJVWFchs7m9mBZnZItvO0eOfiLOArmactTSfcyTghp1rUxDi8rEQ8IiIiItJDvLJ8DcvfXs9Wg/uXL9zgGjZJiHcITgEOBi40s35m1sfMTo/jz4oNiiE8dehc4JPAUXnLM7OdgW3jn7vkTO8LPAjMBP5iZnvH5GNv4A/At9z91kyMLYRkYGvgcjMbFOc5Dvgq8B13b27/VhARERGR7iKparT3tlvUOZLaNWySAODuM4BDCT0WzwXmAKcBx7r7damiC4C7CXX/Z6aXYWZjzOyROH/yea8zs5fM7OzUujYC5wH3AQcS2hvMAE4CTnT3K4vEODvGNzjGNwu4CDjF3b/Z/k8vIiIiIt1Ja5IwuvsnCY3aT0Ird58FnFCmzAbgmCLTFgGHVLiuq4Gr2xHjC4S7GCIiIiLSSyXtEfYevXmdI6ldQ99JEBERERHpLnrSnQQlCSIiIiIiNVq9bgMvvL6Kfn2M3UcNrXc4NVOSICIiIiJSo6dfXYE77DpyKAP79a13ODVTkiAiIiIiUqOe1B4BlCSIiIiIiNSsJ7VHACUJIiIiIiI1U5IgIiIiIiIF5r+aVDdSkiAiIiIiIsDKtRsYMXQgIzcfWO9QOoSSBBERERGRDtBTGi2DkgQRERERkQ7RU6oagZIEEREREZEOoTsJIiIiIiJSQHcSREREREQaRNOUaTRNmVbXGAb07cOuI4fWNYaOpCRBRERERKRGu20zlP59e86pdc/5JCIiIiIiddKTqhqBkgQRERER6aG6shpST2q0DNCv3gGIiIiIiHQXa9ZvzB2/Tw+7k6AkQURERESkhMWvr259v9c3pueW6WnVjZQkiIiIiIiUcMU9z7a+H9CvsLb+ug0tAAwbMqBLY+psShJERERERIpY8ubb/PZvS1r/fubifymYXu9Hr3YWNVwWERERESni6nsXsKHF6x1Gl1OSICIiIiKS459vreHmR1+sdxh1oSRBRERERCTHtfc/z7oNLRw/dtt6h9LllCSIiIiIiGS8sWodv/jrYgDOOXq3OkfT9ZQkiIiIiIhk/PSBF3h7/Ubes+dIxm2/Zb3D6XJKEkRERESkW+jKHpR//tBCAM45evcuWV+jUZIgIiIiIpKxYu0GDttlOAeNGVbvUOqiqn4SzGwscCZwBDCakGQsBWYB9wC/cvdVHR2kiIiIiEhX+0IvbIuQqChJMLM+wPeALwIWX4nRwDjgE8CPzewy4FJ3X73JgkREREREuoEDd9qKw3YdXu8w6qbSOwmXA58nJAdvAw8Di4G+hCThQGBrYAhwIXCqmZ3o7rM6PGIRERERkU6wZv3G1vdfOHp3zKxE6Z6tbJJgZocCZwMO/AT4uru/mVPu3cBZwMnArsADZvY+d3+oY0MWEREREel4f5j1cuv7CXuOrGMk9VdJw+XPxeEv3P3zeQkCgLvf7+4fB94JPEe4q3CHmY1JlzOzrWsJuDsxs0Fmdky94xARERGR8v62aFnr+958FwEqSxLeRbiL8J1KFujujwEHADOBrYCfJtPM7IPAnGoCNLNxZnaLmc03s3lmdk97T7zN7FAzm2ZmR5YoY2b2cTN72MxWmNkSM/u9me1XZtnNZubpF6Fq1uj2xCoiIiIiXeupl9+qdwgNo5IkYTSwyt2fq3Sh8QlHHwZeBSaY2YfN7HrgNmBUpcsxs2OBR4CFwH7AWOBXwAwzO7OK5RxmZtOBvwLvp/Tn/h5wI9BEaIOxXfwsf4tJTt7y3wkcAizJvB4Bfl1pnCIiIiJSH+s3tvD0P1bUO4yGUUmSsKHCcgXc/Z/Ajwkn2r8FJsX3FZ00m9mWhIRgLnChu69z9xZCu4jbgKvMbNcKlrMTMBiYDKwvU/ZwQhIxzt1HA1sAxwCLCI20bzCzwTmzXgCc7u47ZF6HuvvaSj6viIiIiNTPgtdWsm5DS73DaBiVnPwvATYzs23asfxfxKEBrwMT3f2UCuedBAwDbnb3DclId3fgJkKj63PKLcTdF7v7Xe6+GHilTPHzgRPc/ak4b4u73w1MjNNHEKpStTKzvYF9gFsr+VAiIiIi0nieWqKqRmmVJAmPxeFHql24u78ErCW0adjH3Su9i2C0JQB35RS5Ow7PMLMh1YRUYp39gafc/ZlNZnKfSajyBJu2MfgKsAewwMyuN7OTzKxvFTGJiIiISNQ0ZRpNU6Z1+XrnvqIkIa2SJOFGwp2AL5nZZu1Yx1pgjbsvrWKebYCki7unsxPjsl4nVAfasx0xbcLd17v7N0oUSZ7qND8ZYWabAyfFP8cAZxCqSD1pZu/riLhEREREpPM99fLyeofQUCpJEmYQGuDuDvzSqngelJltTziR/0eVcSWPTV1PeEJQnmWZsp0m3mXYDZgNPJWMd/cVwJaExs2nAb8ntOEYB0w3s090dmwiIiIiUht3Z66ebFSgbGdq7u7xSUKPAB8Cms3sE7EqUTlfjMOHq4xrpzhcFtsg5OmyJAH4ADAUuCAbT2xM/QrhjsuNZjYWuIXwJKYrzOz38WlPBcxsMqExNaNGjaK5ubnmIFeuXNkhy5HuQfu799E+7120v3sX7e9Nldoexaa1d/xrq1t4a80GNh8AK9Z1zjoqHd8oyiYJAO4+x8w+CvyO0G/CHDP7MXC1uy/Jlo93G84DvkxoB3BDlXHF3UOpJubJtE7t6SK2L/ga8BN3n16uvLs/ZWbvIdx12BZ4B7BJr9Pufg1wDcD48eN9woQJNcfa3NxMRyxHugft795H+7x30f7uXbS/U6aH9gi526PYtBrHT5/zKvA4+48Zwf3PLu2UdZQd32AqfrRpPEH+F8JV8y2AC4FFZvaYmV1jZheZ2bdifwgLge8TH3/q7n+pMq7FcbhViTLD4nBRlcuu1ucJdy3OrXQGd3+N0N8CZJ6GJCIiIiKNZW5sjzB2uy3rHEnjqOhOQsLdm2N1mu8QGukOBg5k0xPh5Or+b4BPtyOuJEkYZGaD3H1NTplhmbIdLvbMfALwYXcv2cdCjuTuQf+OjUpEREREOlLS0/I+221R50gaR3s6SVvu7ucS2gJ8AbgdeIlQRWgDoV+FW4Dj3f0kd1/djriWAfPi+6bsxNjR2nBgBamnDXWkmAx9hdBvwsp2LCJJKhZ2WFAiIiIi0uGSx5+OVZLQqqo7CWnu/jpwZXx1qNhY+oq47GPZNBGYQEhwbmjnCXxJZrY7cAlwWnyCUXqaAXu4+yaPZs14B7AKqLaqlYiIiIh0kTdWreOV5WsYPKAvTcOr6X6rZ6v6TkIXmgq8AUw0s2ycEwl3La6AcOJuZpeb2VQzG1himUnVn6KNnc2sCfgfYJK7L89M60toi7FL/HsvMzvSzPplyvUjNNr+z85IYkRERESkYyT9I+y17eb07dOpz8PpVho2SYgn16cABwMXmlk/M+tjZqfH8We5+7Ox+B6EhsWfBI7KW56Z7Ux42hDEk/ycMtsTeng+GHjYzOanX4QO1c4E/hxnuQ24F3jEzN4Zk5Udgd8S+ky4rIZNICIiIiKdLOkfQY2WC7W7ulFXcPcZZnYo8G1gLuHuwSvAse5+d6roAuBuQsdmM9PLMLMxwK+BfWlLiq4zs28D33X3K2O5oYQEITeBSLnU3TfG918HPkeoWjQdeBJ4DDjH3TutQbWIiIiIdIynXlZ7hDwNnSQAuPsswhOGSpXZABxTZNoi4JAK1rMS2KvK2H5DeIKTiIiIiHRDSXUjPdmoUMNWNxIRERER6Uyr123g+aWr6NvH2GPU5vUOp6E0/J0EEREREZHOMP/VFbjD7qOGMqh/3w5d9sJLP9Chy+tqupMgIiIiIr1Saydqo1XVKKvT7yTEPgc+BywHfuruL3X2OkVEREREypmrnpaL6tQ7CWY2ALgPOI/wVKLrzSy3gbGIiIiISFeaGxst6/Gnm+rs6kZDgVHAG+5+LfAh4HNmtk0nr1dEREREpKT5r64AVN0oT6dWN3L3N8zsDGBF/HutmX0aWN+Z6xURERERKWfthhZ2GLYZWw7uX+9QGk6nt0lw959l/l7d2esUEREREamEOlHL1ynVjWJbBBERERGRhrbPaLVHyFNzkmBmfczsY2Z2UWr0Dmb2CzN7b63LFxERERHpLLqTkK+mJMHMhgEPAr8CTkvGu/vzwOeBS8zs/JoiFBERERHpJGO3V5KQp9Y2Cd8BDgVeA9amJ7j7W2b2GeAJM3vc3e+pcV0iIiIiIh1m6yED2HaLQfUOoyHVWt3oeOAzwLbAq9mJ7j4b+CdwYY3rERERERHpUPuM3gIzq3cYDanWOwlL3f0GADPzImUGAgfUuB4RERERkQ6l9gjF1Xon4bFSE83saGArYGWN6xERERGRHqRpyjSapkyrawxjt9eTjYqpNUlYYGZj8yaY2VHAzwEHHq1xPSIiIiIiNXti8bLW98fstU0dI2lstVY3uhz4rZnNA0ab2ZeBnYH9gCMAI9xFmFLjekREREREanb9Ay+0vh8ysNP7Fe62atoy7t5iZqcC5wM7At/LFHkY+JK7P1fLekREREREarXkzbe5c84mz9qRHDWnT+7+NvAdM/sesDuwB/AW8Jy7v1ByZhERERGRLjL1oYVsbCn2rB1JqylJMLMx7r4IwN3XAH+Pr3SZo4GH4nQRERERkS63au0GfvnI4nqH0W3U2nD5hgrKPAB8scb1iIiIiIi0262Pv8SKNRs4aMyweofSLdSaJFRiIPD+LliPiIiIiMgmWlqcGx4MteA/866d6xxN91BVdSMLXdJ9DzgwjtrfzO4uVhwYSWinsLzdEYqIiIiI1OCu+f9k4eur2X6rzThun1H1DqdbqCpJcHc3s/8ArgROJ/SBMKHMbG8D32hXdCIiIiIiNbr+gecBOP2IJvr17YqKNN1f1Q2XYwPkz5jZAuAk4LwSxVcAz7r7inbGJyIiIiJSk78+/wZDBvTl5IN3rHco3Ua7n27k7peY2Rx3v7dUOTM7yMyedveV7V2XiIiIiEgtTj54R7YY1L/eYXQbNd1vcfc/VFBsOPCjWtYjIiIiItJeZnD64WqwXI2aO1Mzsz0J1Y62AwZlJg8ADgK2Ac6sdV0iIiIiItU6bp9R7DR8cL3D6FZq7UztA0ByN8EIDZktp+j9taxHRERERKQa9z7zWuv7z7xrlzpG0j3VeifhP4EZwG3AauDjwC9T00cAhwGn1rgeEREREWkATVOmAbDw0g/UOZLi7nvmNT479bHWvw9uUgdq1ao1SRgMHObuGwHMbADwsLs/kxQws0uAscCTNa5LRERERKSkB55dymenPsa6DS2t40JXX52jkZOlWtT6oNjXkwQh+iVwdqbMDOCbNa5HRERERKSkB59bymd+/ihrN7Rw6iE71TucXAsv/QA/O35IvcMoq9Yk4W0zm2RmB5jZAHdfBWw0s0mpMu8HjmjPws1snJndYmbzzWyemd1jZse0c1mHmtk0MzuyTLmdzOw6M3vGzP5uZjPN7KQS5bc2s8ti+SfNbLaZfc46M2UVERERkU0kCcIpB+/If50wrt7hdGu1Vje6DLiT0GD5d8CJhN6YnzSzbxIaMe8EvFTtgs3sWEKj6P8FPglsAP4NmGFmk939ugqXcxih7cT74qjvlSi7H3A3MB04gNDO4oPAbWa2r7t/M1N+e+AB4AXgUHdfZmYHA/cCh5nZJHf3Sj+ziIiIiJT3wLNL6dvH6NfX6Nun7brsmvUtnDx+By75yL706aPrtbWoKUlw9z+Z2acIVYzujuMWmNlpwK1AX0IC8e1qlmtmWwK/AuYCF7r7hjj+J8BxwFVmdo+7LyiznJ0I7SYmA88BRXvQMLO+Mea3gbPiXRGA283scuAbZvYXd78vNdvPCY93fZe7L4uf/1Ez+xqhb4i7gKnVfHYRERGRemvExskPLVja+v6062fmljnxoB249KP7KUHoALVWNwK4290Pc/f/S0a4+23A7sBEYJy7/7TKZU4ChgE3JwlCXK4DNxGSm3PKLcTdF7v7Xe6+GHilTPEPxphvc/cVmWk3xeGXkhFmtj9wDHCPuy/JlL85Ka9qRyIiIiK1u+2Jl1vfH77rcA7deWvGjxnG/jtu1Tr+vz+mBKGj1NpPwp3AcWZ2gbt/Pz3N3RcCC9uxTKMtAbgrp8jdcXiGmf1H6op/OeWq/XyhxDofB94EPmRmY9x9UakY3f0fZvZ3YH9Ce4wHKoxRRERERDI2bGzhz3Nfbf37l599Z8H05M5H3w5IEBrp7kk91XonYXwczi5VyMyqSUa2AXaL75/OTnT3pcDrwBbAnlUst6iYmBxeYp0bgQWE7XVQHH1EsfJR8hjYQzoiRhEREZHe6pEX3mDZ6vX1DqNXqTVJmAKsBB4qU66aHpfHxOF6QvuAPMsyZWs1Etgsvn+z3DpjUjGm0vK1hyciIiLSe90559XyhaRD1dpw+Xozexa4xMy+6+6b1Ps3s32B/apYbPJQ22UlngzU0Sfg6QfpLitSJr3O4bQlFZWU34SZTSY0qGbUqFE0NzdXGmtRK1eu7JDlSPeg/d37aJ/3LtrfvUu99vek6aHWdrHn9peKqdJ427uOZHyLO7c/8XbutGqXVen4UnrLOVutbRJmAclDaM/uoDa66+KwpUSZZFpHtUxZl3pfbL3pda7PGV+q/Cbc/RrgGoDx48f7hAkTKgq0lObmZjpiOdI9aH/3PtrnvYv2d+9St/09PdTl32TdxcaXm9YR68iMf3zRG7z5p4fZfqvNWPLm2zUtq92fob3zFNEdvt+1Vje6knASXO5VjcVxuFWJMsPicFGVyy63zlLrbV2nuy8HlldavsbYRERERHqtO/8eqhodP27bOkfSu9SaJPwCeBHY3d375L1ou9NQqeSEfZCZDSpSZlimbK2WE9pWQPmT/sWZYaXlRURERDpV05RprU/66QncvbU9wr8oSehSNSUJ7r4auKBUp2buPpdwx6FSy4B58X1TdmLsaG04sAKYX8Vyi4ptH5LG1TvnrLNPHO+Ex6FC22NNNykf7RKHj3REjCIiIiK9zZwlb7HkzbfZZvOBHLjTsPIzSIepuTM1d7+5gjLnVrE8B66Ifx6bU2QCIe4b3H1lzvT2KrXOfYERwO3u/kIcd2Wx8ma2NXAg4dGw1TzZSURERKSsnnbHoJg754Rn4rxv7LbqJK2LdUSPy51hKvAGMDFexU+bCGwgntRbcLmZTTWzgSWW2T8Oix1h0wl3Jk7IqeY0MQ4vS0a4+1PAn4FjzWxEpvyJcT0/LPGEJhEREREpwt2ZrqpGddOQSUK8Q3AKcDBwoZn1M7M+ZnZ6HH+Wuz8bi+8BnAt8Ejgqb3lmtjOQHF275JVx9xZCMrA1cLmZDYoJyHHAV4HvuHtzZrYzCO0Zfm5mm8d17Q/8gJDoTK36w4uIiIgIz/xjJc8vXcWwwf05ZOet6x1Or9OQSQKAu88ADiX0cDwXmAOcBhzr7telii4A7ia0FZiZXoaZjTGzR+L8yWe9zsxeMrOzc9Y5O65vcFzfLOAi4BR3/2ZO+SWx/ELgCTN7Avg5cD4wSXcRRERERNonqWr03n1G0a9vw56y9lg19ZPQ2dx9FnBCmTIbgGOKTFsEHFLlOl8g3JWotPxSYJOEQ0RERETar62q0eg6R9I7KS0TERERkYYz/9UVbD6wH4fvNrzeofRKNScJZrZdRwQiIiIiIpJ2zN7bMLBf33qH0SvVlCSY2Z3Ai2Z2fgfFIyIiIiIdrLs+MvV4VTWqm1rvJIyPw9mlCplZQ7d9EBEREZHGsln/vhy1x8h6h9Fr1ZokTAFWAg+VKacOxURERESkYu/ZaySbDVBVo3qpKUlw9+uBfwUuMbPc+0Fmti+wXy3rEREREZHeZcKe29Q7hF6tpmpAZjYLGBf/PNtM3WWLiIiISPuku5g6pEkdqNVTrdWNrgSsgpeIiIiIVKG7NTbuiHifX7qq9f2Y4YNrDUlqUGuS8AvgRWAPd++T96LtToOIiIiISFGPLXyj9b1qqNRXrW0SVgMXuPtzAGY2yswOMLPNUmXmEu44iIiIiIgU9ejCZfUOQaKaH03q7jeb2WTgW8CoZLSZPQH8u7vf7+7n1roeEREREenZ0ncSGs3CSz9Q7xC6VEf0uDwVuArYltD+YAOwBjgIaDazC2pdh4iIiEh3V6zOftOUaUyavipnjt7lnyvWsPD11fUOQ6Jae1z+KHAaMBP4BLA9MMjdhwIjgLOAc8zssFoDFREREZGe63FVNWootVY3OgO4Hpjs6WdWAe7+BnCNmf0VuBB4uMZ1iYiIiEgPpfYIjaXWJGFP4GPZBCHN3Web2dAa1yMiIiIiPdhjixq3PUJvVGubhBfcfW2pAmY2iJBMiIiIiPQY3a0fg0a2au0Gnnr5Lfr20WNPG0WtScIaMyvaZ7aFB9xeAjxX43pEREREpId6YvGbbGxxxm63Rb1DkajWJOF24HYzK+gwzcwGmtmJwN3AF4Ff1rgeEREREemhHo2PPh0/Zus6RyKJWjtTuxZYDswys5fM7G9mtgRYBdwCHAXc5u5Taw9VRERERHqipD3CwU3D6hyJJGruJwH4V+BiYCSwPzA6LncNMAU4uQPWISIiIiI90PqNLTyx+E0AxjfpTkKj6Igel9cC/2lmlwJjgSZgATC3XKNmEREREend5r3yFqvXbWTnEUMYufnAeocjUc1Jgplt5+4vu/vbwGPxJSIiIiJSVtI/wvgxqmrUSGrtcflO4EUzO7+D4hERERGRXuSxhUl7BFU1aiS1tkkYH4ezSxUys5rvWIiIiIhIz9N6J0GNlhtKrUnCFGAl8FCZcvfXuB4RERER6YGWrlzL8CED2HnEkHqHIim1PgL1esLTjS4xs9F5ZcxsX2C/WtYjIiIi0pnUe3J9jW8aRuiDVxpFTdWAzGwWkHSkdrZ2roiIiIhUS+0RGk+t1Y2uBKyCl4iIiIhILvWP0HhqTRJ+AbwI7IVHB8oAACAASURBVOHuffJetN1pEBEREREpMKh/H8Zut0W9w5CMWtskrAa+7u7PlSgzl3DHQURERESkwAE7DqN/31qvW0tH64h+EqaW6yfB3c+tZT0iIiIi0jMdrEefNiT1k9CJzGyomR1Z7zhEREREGklLi7e+P3SX4XWMRIpp+H4SzGycmd1iZvPNbJ6Z3WNmx3TGMszsIDPzMq9NEiIz62dmC7JlgRVAS3s/u4iIiEg9vbp8Tacs98EFS1vfv1NJQkNq6H4SzOxY4BFgYVzGWOBXwAwzO7MTljGZcFK/EHg683o2lvlDzmpOBEYCSzKvO4AHK4lTREREOp/6Q6jOlfcUbXZakxv/urj1fd8+ehBmI2rYfhLMbEvCyfxc4EJ33xDH/wQ4DrjKzO5x9wUdsQwz2xw4Btjb3Z/JWda7gfuA32TGG3AB8F53n1nr55b/396dx8lR1/kff31mJsmE3HdIyEUOQEISQrivABEPkGUFJN5cIij680aWZRXc3y7qqssuKCA/QZQFFBEE1giajCAC4UpCICF3QkJCTkIm5xyf3x/17Zmenu6e7p7u6e6Z9/PxqEd1V33r29+u6qmpT9X3EBERkWJ6ee32ptePLXqbfz73CHpUVeYt/0079/HUknfylp8URimPk3AJMAB4IHZxD+DuDtxPFOBck8c8JgOfTRYgBB8DVgELEpafDWxXgCAiIiKdwc+fXt30+t09dcxbujmv+T/44ls0xLVJkNKUr3ESJuZznIRwdz528f6XJEnmhvllZtYrH3m4+3PunrRqkJlVElUpeigEGPGuBc4MbR1uM7MPmoaeFhERkTK0Zutu/vTGphbLfvfKhrzlX9/QyAMvrms7oRRdPsZJuDZZlR8zO8XMDs5xnIShwITw+s0kn7sV2Ab0BQ4rYB4xpwDDaV3VaAIQ673ocOALwB+BZ8zs+DbyFBERESkpd/1tFfG3QysrjHlLN7Otdn9e8p/35hY27tzHuMFJ7/FKCck4SDCzKXHTsNhyd38gxSYDgFfM7BVgZ5blGhPmdcDeFGl2JKQtRB4xHyN6YvJi/MIwiFyPsP2VwFNh1cnA02Y2s418RURERErCttr9/Pal9S2WnTZxMPWNzmML387LZ/z6+bUAfPL40XnJL1drbj6HNTefU9QylLpsGi4vAPYTNQS+HUjb4sTdHzOzo4GXgH8Cbsjis2K/nB1JqvfEtHWBn488YlWNLgD+J1k+7t4ArAN+DvzczE4i2kcjgTvM7PBk25nZlUSBBcOGDaOmpiZVETJWW1ubl3ykPOh4dz065l2Ljndhpdq3uezzbLfJ9rPT5Z+vvGpqanh0xQH21zcyZUgli7Y0AHBE9XvMA+7561LG1q1t12f85n/n8vSyvVRVwPC9a9tM3xHfu1jK4e87296NZsXX2zezB4BJcev3A2+6+yUA7r7JzL4F/CrLzzkQ5unGGIitS1X/Px95AJwKDCOhqlEq7v53MzuDaIC5ScDBQKvw293vBO4EmDFjhs+cOTOT7NOqqakhH/lIedDx7np0zLsWHe8CmRN1f9pq36ZankNesS5WW92pzvaz05UpX3mF5SecfCpffyZqrnnd+cfy8Z8/D8CXLzyDX735Z9a8V8+II45h0rA+OX/G6ooROCs5b+pIzj17GtfMLeL3LrJy+PvOpk3CqsSGve4+G/gBMA14BDgjFiDEmZNDuWItWvqnSRMbw3ttivX5yAOiqkYbgefSpGnB3ZcTPVkAODrT7URERESK4XevrGfb7gMcNbIfJxw6sGl5dbdKzp0yAoCH29mA+TcvvgXAJ09oq5a3lIJsgoQtyRaGNgmb3f0md281LJ+7bweyHa4vdoFfbWbVKdIMSEib9zzMrIqoqtHD7p7tyMmxUai7ZbmdiIiISIe665mo29PPnXYoiZ00XjB9JAC/f3V9u7ou3bb7AIcP78P00enu3+aX2h7kLpsgIVXjX4BUYwvE7MricyBqK7AkvB6buDIMkjYo5Lu0gHmcRtRL0kOZFbuFujBfk8O2IiIiIh1m9dbdjOzfkw9PHt5q3TFjBjBm0EG8895+/r5ya7s+55MnjGkVhEhpau84CTH1baxvyCaz0ND31vB2VpIkM4nKfre71xYqD6KqRluAZzIqeEtTiQKEhTlsKyIiItKhLj9lHFWVrS8NzYyPHn0IAL97eX2r9Znq1b2Sfzx6ZM7bS8fKV5DQVkg4sI31ydwLbAcuNrPEcl5MFJjcCtHAaWZ2i5nda2Y9cskjUVxVo9+HHoySpZluZsclDp4WnlJcDXwzTc9KIiIiInnR0Oj86vl0TSzT61tdxcXHjkq5/qOhytGc1zelTNOWfzh6JL17ZNtnTkuqPtRxsgkShiVcgMdLeSFsZuOA7lmVCgh392cDxwLXm1mVmVWY2aVh+dWhgTBEvQh9Gfg0cHqOeSSaCQwmRVWj0DXqs8ALwJ/N7IgQrBwB/AH4rrvnUk1JREREJGML3nqX8297lhseWdy07EB9dk0pP3XCGHqluYAfNfAgjhs3kH112eW7r675Pusnjivu2AiSnWzCuSOAPanqkZlZVlWKMuHuT4WRi28E3iC687+RqCvWuXFJVwJzgX5EF+255JHoY0TtGmpSlK3BzL4CfAKYTlQlaQHwNHChuydt6C0iIiKSL9c9/BoPvLgOdzi4XzUbd0Z9xSx4612OG5e+Ikd8IJFJj0MXTB/J/NXbsyrf3KWbm15PHtkvq22luLJ95pNrS5Ocq9y4+0Lg/DbS1ANntSePJNs0DXaWJs0dwB3Z5CsiIiLSHvE1me+fv46qCuPy08bx5TMncuR3/gTA31ZsbTNIeHntjqbXI/v3bPNzP3zUwfzLo6+zP4unFO1pwyDFlU2QUA88QDSmQFsNlePzPxS4KMtyiYiIiEgSv4278D7x0EF87/wjmTC0T4s0z67YytfePylx0xaeXp5dpYc+1d34wJHD+cPCVmPEJrW1dj81y1SxolxlEyT8wt2vyuVDzCzbMQZEREREJIG78/OnVzW9/5/PHZ+0S9EFb73Lrn119KlOPVzT0zlcwH9sxqimIGFfXQPV3SpTpv3DgrfbNa6CFFc2DZcfaMfnqAGviIiISDv9bcVWlm9u7rk9VVvRhkbn+VWp2w9s2bWf199+L+vPP3nCoKbXv381/QjMD7+qqkblLOMgwd1rcv0Qd380121FREREJHL3s2syTvvsitQDnz2TZVWjmPig5M6nV6V8UvDmpl0s3vAefavb1+WpFE++xkkQERERkQJavXU3c5dupntVZpdvf0sTJORS1ShZeZ5MMW7Cw69ETxHOnTqi3Z8jxaEgQURERKQM/PLvawA4f1rbF969uleyYnMtm0KXqPEaG51nlqcOILJx+19XkjhubEOjN1VFumC6RlguVwoSRERERErce/vq+O1LbwFw6cnj2kx//KFR24FkVY7e2Pge23YfYES/6naVaWCv7ixcv5PnVm1rsfzZFVvZvGs/YwcdxPTRA9r1GVI8ChJERESk0xn77ScY++0nil2MvPntS+vZfaCBEw4dyBEH920z/ckTBgPJg4S/hqpGp00a0q4yffbEsQDc/tdVLZbHqhp9dPohKRtWS+lTkCAiIiJSwhoavamqUSZPEQBOCUHC31ZsbVUd6Ok8BQmfOXEMPbtV8vSyLbz+9s6m5XNCO4V/PFpVjcqZggQRERGRAtu1r67p9W3zVtCYxfgB85ZuZt32PYwa2JNZRwzLaJtJw3ozpE8PNu/a36LL1Nr99by8dgcVBiePH5z5F0hiQK/uzD5uFBD1dBSzr66R48YNZNTAg9qVfzbW3HwOa24+p8M+rytQkCAiIiJSYPFjEvzwT29y1a9fbhE4pHP331cDUfWeyorMqu+YWfPThLhGys+t3EZ9ozNtVH/6HZR6oLVMXXHqoVRVGI8v2thiuRoslz8FCSIiIiIp5Kttw+INzdVx+lRX8eQb7/APtz7L8nd2tbntsyu2cVD3Si6aMSqrz0zWLiFfVY1iRvbvyXlTR7QYL6FHVQUfOurgvOQvxaMgQURERKTA4p8kPHbNKRw+vA+rtu7m/NuezWj7C6YfQr+e2d35j42O/Hxc70P5arQc7/Onj2/x/gNHDqdvdfufUkhxKUgQERERKbD4JwljB/fi4S+cxHlTR7D7QEPT8vteWMtTb7zDwrfeZePOvS22v+TksVl/5sH9ejJ+SK8Wn7Fu+x769ezG1EP6Z/8lUjhseB/OPHxo0/uPqqpRp6CxskVEREQKaO+BBlZuqW2x7KDuVdwyexpTR/Xne4+/AcD1v1+cdPvTJw1h/JDeOX32KRMGs3LL7pbLJg7OuG1Dpq46fTxzl25u+kwpf3qSICIiIlJASza9R7LOjMyMy09p7tL04hmjOOOwIUwe2ZehfXo0Lf/iGRNy/uyTk1ywnz4xf1WNYo4bN7DpdVWlLi87Az1JEBERESmg1+OqGqXz/QuntHgfazAdfwGerRPGD6LCaBGknDpJd/qlbQr1RERERApo8Yb32k5UIH2ruzF1VHP7g0nDenNwv55FK4+UDwUJIiIiIgW0+O3MniQUyqlxVY5OK0BVI+mcFCSIiIiIFMiB+kaWvbMLy2874azEt0vIZ9en0rkpSBAREREpkGXv7KKuwRk3uFfRynD06AFNr9vTvkG6FgUJIiIiIgXyeqhqNHlEv6KVoXtV8+VedbfKopVDyouCBBEREZECiTVanjyyb5FLIpIddYEqIiIiUiCxRstHFvFJQrlZc/M5xS6CoCcJIiIiIgXR0Ogs2Rg9SThyhJ4kSHlRkCAiIiJSAKu21LKvrpFDBvSk/0Hdi10ckawoSBAREREpgMUl0GhZJFcKEkREREQKQI2WpZwpSBAREREpgMUbQqPlkXqSIOVHQYKIiIhIAbzxthotS/lSkCAiIiJdwt4DDU2v6xsa25XXzr11babZtb+eoX16MLRPdbs+S6QYFCRkwcx6m9lpxS6HiIiIZO+BF9c1vf7RU8valdePn3yz6bW7p0w3WVWNpEyVfJBgZpPN7EEzW2pmS8xsnpmdVcg8zKzKzFaamcdPwC6g1a0HMxttZneZ2TIze83MXjCzi7L/tiIiIlIIdQ2N/PzpVU3vf1azkjmLN+WU154D9Tz8yoam948s2JAy7WRVNZIyVdJBgpnNAuYDa4ApwJHAb4CnzOyKAuZxITAE2JAwPQ48m5D/FOAVoBo4OnzGvwIPmtlNmX1TERERKaQ/LHibt3fua7HsG79dyKottVnn9djCt9m1v77p/b8+voR39xxImrZcGy2vufkcjXzcxZVskGBm/Ygu5t8Arnf3A+7eCNwOPAr8zMzG5zsPMzPgWuD97n5IwvQRj3umaGaVwEPAXuBqd9/tkceAW4AbVD1JRESkuBobndv/urLFsnOOOpja/fVc9euX2R13wZ+J+15Y1+L9tt0H+P6cN5OmVaNlKVclGyQAlwADgAfcvemvN1yk3w9UAdcUII+zge3u/kIGZTwXmAg86u67EtbdH+ZfzSAfERERKZC/LN3M8s21jOjX3ID4+xdOYcLQ3ix7p5Zrf7co47wWrX+XRet30v+gbk3LulUa989fx8trt7dI2/+gbozs37P9X0CkCEoySAh382MX739JkmRumF9mZr3ynMe1wJmh7cJtZvbBkFcyX0qT/8vAu8B5ZjYmxfYiIiJSQO7OT2tWAHDFqYc2Le/do4rbP3UMvbpX8viijRnnd9/z0VOEC6cf0rTsytOifK///WLq4npNmjyiH6kvIURKW0kGCcBQYEJ43er5nbtvBbYBfYHD8pWHmU0AYtWDDge+APwReMbMjo/fPgQOJ6XJvwFYSbSPj0lRRhERESmg+au38+q6d+l/UDdmHzeqxboJQ3vzHxdNzTivnXvr+MPCtwH4+PGjm5Zfc8ZERg3sydJNu7jn2TVNy4/USMtSxko1SIjdea8jqu+fzI6EtO3Ow91XAD3C+yuBp8L6k4GnzWxm3LZDgNgzxHdzLKOIiIgU0E9rorYIl5w0loO6V7Va/6GjDubzpzU/YXjnvX2t0sQ88uoG9tY1cNL4QYwf0rtpec/uldx03mQAfhzXterkEeXZaFkEojr5pSgWnu/w1J0Pt3UBnlMe4QnAOuDnwM/N7CSixs8jgTvM7PCQ3+gk+WRcRjO7kigQYdiwYdTU1KTIInO1tbV5yUfKg45316Nj3rXoeLftkjm7Abjng0lrHgPw12Vb6F4JExo3UFPzdtPy+H17XLVzR3j98dvmcd3x1XSraFlNaN68edz5bHTPcVqflsempqYGA2YMq+Sld5oHbNu9fik1O1qPx5DquGa7PJ955fJbK+Znl7ty+Psu1SAh1o9YuuEQY+tSVfbLRx64+9/N7AxgETAJOBh4Oy7/dJ+RMn93vxO4E2DGjBk+c+bMNMXMTE1NDfnIR8qDjnfXo2Peteh4Z2DOEwDJ91NYB/CpE8Zx7tnvS7/Nk9HyVTsbeWr7QL5/wZSoPUFI33vcVDb86TkG9+7BVy48k+5VFa3yOvzofZz1oxp2h5GdP/ahM6iIDzZSfXa2y/OZV7rPSKWYn91JlMPfd6lWN4r1LdY/TZoBYb62gHkA4O7LiZ4sQDQWQnz+6T4jo/xFRESkMKoqjCtOHZdx+h5VFfzmpfX8+vmW/7rvC+8vPvaQKEBIYni/ar5+dnNTyYoKNVqW8lXqQUK1mVWnSDMgIW0h8oj39zCP9Xm2E4iNwNJWkJBJ/iIiIpJn5x89khFZdEP6/QumAHDjY28wf3Vzl6b/+9omzGD2saNTbQpEbR9EOoNSDRJ2AEvC67GJK8MgaYOAXcDSAuYRry7M10DTWAvPhGWtblGYWUVY7kTdoYqIiEgH2LyrufHxVacfmiZla+cfPZIrThlHfaPzhfua/30faGhk5qQhjBp4UNrt9fRAOouSDBLCBfit4e2sJElmEpX9bndPOp56PvJIMJUoQFgYtyxd/kcBg4HH3H11BvmLiIhIHsxburnp9YShfbLe/tsfOpyTJwxia+2BFss/dYI6K5SuoySDhOBeYDtwcbgrH+9ioJ5wkW6RW8zsXjPrkUseIZ/pZnZc4uBp4anD1cA3E3pKmkP0FOL8JFWaLg7zn2T2dUVERCQfFq7f2a7tqyoruPXj0zlkQHM1pRH9qpl52ND2Fk2kbJRskBDu7s8GjgWuN7MqM6sws0vD8qtDg2KIeh36MvBp4PRc8jCzSuBZ4AXgz2Z2RAg+jgD+AHzX3R9KKGMjUTAwELjFzKrDNmcD3wK+5+41Bdg9IiIiksKi9amGL8rcgF7dufPTM5ref/y40VSqKpF0IaXaBSoA7v5UGOn4RuANojv/G4FZ7j43LulKYC7Qj+giP+s83L3BzL4CfAKYTtTeYAHwNHChu29JUcZFZnYMcBOwGNgTptmJQYWIiIgU1r66BpZu3JWXvN43onnE5M+ePDYveYqUi5IOEgDcfSFwfhtp6oGz2pNHSHcHNI2nkk0ZVxM9xRAREZEiWrLxPeobU42hmru+1d3aTiTSiZR8kCAiIiKSqUXtbI8gHWfNzecUuwiShoIEERER6TS6epCgC2/Jl5JtuCwiIiKSrXw0WhYRBQkiIiLSSdTur2fFllqq1AuRSLspSBAREZFOYfGGnbjD4QdnP4CaiLSkNgkiIiLSKcSqGh01sj+LN7xX5NJIjNpJlCcFCSIiItIpxEZannpIP+6fX+TClAldwEsqChJERESkU4g9SZhySP8il6TwdHEvhaYgQURERMrejt0HeGv7Xqq7VTBpWO9iFycpXdhLOVHDZRERESl7izZEVY2OHNGPqkpd3oi0l/6KREREpOwteitW1ahfkUsi0jkoSBAREZGyF2u0rCBBJD8UJIiIiEjZ60qNlkU6goIEERERKWubdu5j86799OlRxbhBvYpdHJFOQUGCiIiIlLWmQdQO6UdFhRW5NCKdg4IEERERKWuLmtojqKqRSL4oSBAREZGytjA8SZiqRssieaMgQURERMraa2GMhKMUJIjkjYIEERERKWvv7qljUK/ujOzfs9hFycmam8/hng+qwbWUFgUJIiIiUvamHNIPMzVaFskXBQkiIiJS9tRoWSS/FCSIiIhI2Zs6Su0RRPKpqtgFEBEREWmvo0bqSUJHWXPzOVktl/KkJwkiIiJS1kb0q2ZInx7FLoZIp6InCSIiIlLWOmt7BN2Zl2LSkwQREREpaWO//QRjv/1EyvVTR3XOIEGkmBQkiIiISFk7/tCBxS6CSKejIEFERETKzqad+5peHzVSPRuJ5JuCBBERESk7z63a2vS6W6UuZ0TyTQ2XRUREpOw8t3JbTtupMbBIZhR6i4iISNl5ftX2YhdBpFNTkFBAZlZtZmcVuxwiIiKdyYZ397Ju+55iF0OkUyv5IMHMJpvZg2a21MyWmNm8bC+8s8nDIp8ws+fMbJeZbTCzR8xsShufUWNmHj8Be4GDsymriIiIpJdJVaM1N5+jqkUi7VDSQYKZzQLmA2uAKcCRwG+Ap8zsigLl8UPgPmAsYMAI4B+AV8zs3BSfcQJwHLAhYZoP/DaTcoqIiEhmcm2PICKZK9mGy2bWj+hi/g3genevD8tvB84GfmZm89x9Zb7yMLOTgA8Dk939dTOrAGYCvwDGAHeb2Rh3T3zGeS1wqbs/mKevLyIiIkm4O8+v6rggQU8jpKsq5ScJlwADgAdiF/cA7u7A/UQBzjV5zuMbwPnu/npI1+juc4GLw/rBwNHxH2BmRwDvAx7K7uuJiIhItt7avpcN7+6lX89uxS6KSKdWkkGCmRnNF+9/SZJkbphfZma98pGHmXUDXnf3ZYkJ3f0FoupK0LqNwTeBScBKM/t/ZnaRmVUmK5OIiIi0T2x8hOPHaZRlkUIqySABGApMCK/fTFzp7luBbUBf4LB85OHude5+Q5oyvRvmS2MLzKwPcFF4Owa4jKh60wIz+0CavERERCQHsfYIJ44fVOSSiHRupRokjAnzOqIegpLZkZC2EHkAEJ4yTAAWAa/Hlrv7LqAfUePmTwGPAPXAZGCOmX0yXb4iIiKSuag9QjQ+goKE7Ki3J8lWqTZcHh3mO0L7gWTausDPRx4x5wC9gWsT83L3RmAjUY9I95nZkcCDRL0o3Wpmj7j77sQMzexK4EqAYcOGUVNT00YR2lZbW5uXfKQ86Hh3PTrmXYuOd2sP/u88Nr23jz7d4O0lLzctz+d+yiWvVNtkk5eOd9dSDse7VIOEA2HemCZNbJ0VMA9C+4LrgNvdfU6avAAIvSKdQfTUYTgwFfh7knR3AncCzJgxw2fOnNlW1m2qqakhH/lIedDx7np0zLsWHe84c54AoHHwBOA1Tj18OGeecQz8KVqel/00J4e8Um2TQ1463l1LORzvUq1utC7M+6dJMyDM1xYwD4AvED1x+HKaNC24+xai8RYgoTckERERyc1zoevTEw9VVSORQiv1IKHazKpTpBmQkDbveZjZacD5wIXuXpemvMnEnh6ojzYREZE8iDVaPkFBgkjBlWp1ox3AEuAIopGPl8avDIOkDQJ2Ja7LVx6hbcE3icZNqM3hO8SCijU5bCsiIiIJttbuZ3DvHkwY2rvYRUlLDYSlMyjJJwmhcfCt4e2sJElmEpX97lQX8O3Jw8wmAv8GfCr0YBS/zswsVber8aYCu4E/Z5BWREREMnDCoQOJhkISkUIqySAhuBfYDlxsZonlvJioq9FboenC/RYzu9fMeuSSR4yZjQV+AFzi7jsT1lUC1wOHhveHm9lpZlaVkK4K+DrwnRyfQoiIiEgS6vpUpGOUbJAQLq5nA8cC15tZlZlVmNmlYfnV7r48JJ9E1LD408DpOeaBmY0kGp35WOA5M1saPxENqHYF8GTY5FHgr8B8MzshBCujgIeJxkz4SSH2jYiISFelRssiHaNU2yQA4O5PmdnxwI3AG0R3/jcCs9x9blzSlcBcooHNXsglDzPrTRQgHNpGsW5294bw+p+Aq4iqFs0BFgAvAde4e6oG1SIiIpKDYX17MG5wr2IXQ6RLKOkgAcDdFxL1MJQuTT1wVjvzqAUOz7JsvwN+l802IiIikpsTDx1UUu0R1EBZOrOSrW4kIiIiEk/tEUQ6joIEERERKQtTR6UbH1VE8klBgoiIiJSsfXUNTa/VHkGk4yhIEBERkZK1asvuptc9qiqLWBKRrqXkGy6LiIhI17V88662E+WBGiGLtKQnCSIiIlKyVm7WmKQixaAgQURERErWcgUJIkWhIEFERERKloIEkeJQkCAiIiIlqa6hkTVbd7edUETyTkGCiIiIlKS123ZT3+jFLoZIl6QgQURERErS8ndU1UikWBQkiIiISElSewSR4tE4CSIiIlKSVrQRJGhsA5HC0ZMEERERKUl6kiBSPAoSREREpOQ0NDortyhIECkWBQkiIiJSctbv2MOB+kaG960udlFEuiQFCSIiIlJyYj0bTRzWu8glEemaFCSIiIhIyYm1R5gwVEGCSDEoSBAREZGSs3zzLgAmDu1T5JKIdE0KEkRERKTkrNSTBJGiUpAgIiIiJcXdm6obTVSQIFIUChJERESkpLy9cx97DjQwuHd3BvTqXuziiHRJChJERESkpCx/J2qPoKpGIsWjIEFERERKygq1RxApOgUJIiIiUlJWNLVHUM9GIsWiIEFERERKihotixRfVbELICIiIhLj7q2qG625+ZxiFkmkS9KTBBERESkZW2r3s3NvHX2rqxjSp0exiyPSZSlIEBERkZKx4p1Q1WhYH8ysyKUR6boUJIiIiEjJUHsEkdKgIEFERERKhro/FSkNChJERESkZCzfrIHUREpBSQcJZjbZzB40s6VmtsTM5pnZWYXMw8xGm9ldZrbMzF4zsxfM7KI06Qea2U9C+gVmtsjMrjJVpBQREcla0xgJwzRGgkgxlWyQYGazgPnAGmAKQf4TlgAAGO5JREFUcCTwG+ApM7uiEHmY2RTgFaAaODps86/Ag2Z2U5L0I4GXganA8e4+Dbgc+DFwjwIFERGRzO3YfYCttQfo1b2SEf2qi10ckS6tJIMEM+tHdDH/BnC9ux9w90bgduBR4GdmNj6feZhZJfAQsBe42t13e+Qx4BbgBjM7LeFjfgkMBT7t7jsA3P1F4DrgM8Cn27UjREREupAVW6KnCOOH9lbPRiJFVpJBAnAJMAB4wN3rYwvd3YH7iQaBuybPeZwLTAQedfddCXndH+ZfjS0ws2nAWcA8d9+QkP6BWHo9TRAREcnM8nfUaFmkVJRckBAuqmMX739JkmRumF9mZr3ymMeX0qR/GXgXOM/MxoRlKfN393eA14BpwMnJyigiIiItxRotTxyq9ggixVZV7AIkMRSYEF6/mbjS3bea2TZgEHAYURuCduVhZq8CJ6VJ32BmK4FjwrSW5ov/VumDZcBRwHHA31KkKZrGRqfRvdjFkHZoaHTqGxqLXQzpQDrmXUtXPN4rNEaCSMkoxSAhdqe+jqh9QDI7iC7wx5A8SMg2j/VAz7D83TTpAcaEJxVjMk2fYn1R/XHxJr74P8l2nZSVJ/9Y7BJIR9Mx71q66PFWdSOR4ivFIGF0mO8I7QeSaesCPNs83kqyPF36QTQHFZmkb8XMrgSuBBg2bBg1NTUpsslcbW1txvm8sameCrWWKG/uoCYvXYuOedfSRY/3hP4VrHptPmu62HfP5n+4lL9yON6lGCQcCPN0z1hj61KdQbLN40CS5enS12WZvhV3vxO4E2DGjBk+c+bMNEXNTE1NDZnmMxP45ux2f6QUUTbHWzoHHfOuRce7a9Hx7lrK4XiXXMNlYF2Y90+TZkCYr81THuvilqfapim9u+8EdmaaPk0ZRERERERKTikHCdVmlmoklQEJadubx06gNrxv66J/XcI80/QiIiIiImWhFIOEHcCS8Hps4sowSNogYBewNB95hHYLz4TV45KkrwjLnag7VGjusahV+uDQMJ+fYr2IiIiISEkquSAhXLDfGt7OSpJkJlG573b32iTrc80jXfqjgMHAY+6+Oiy7LVV6MxsITAcW0Rx8iIiIiIiUhZILEoJ7ge3AxeEufryLgXrCRb1FbjGze82sRy55BHOInkycn6SK0sVh/pPYAnd/HXgSmGVmgxPSX0jUYPnHaXpXEhEREREpSSUZJIS7+7OBY4HrzazKzCrM7NKw/Gp3Xx6STwK+DHwaOD3HPHD3RqJgYCBwi5lVhwDkbOBbwPfcvSahqJcRtWf4pZn1ATCzacCPiIKUe/O3V0REREREOkYpdoEKgLs/ZWbHAzcCbxDd+d8IzHL3uXFJVwJzgX7ACznmEUu/yMyOAW4CFgN7wjTb3R9Kkn5DSH8j8KqZ7SIKvL4B3KmnCCIiIiJSjko2SABw94XA+W2kqQfOak8eCelXEz2VyDT9VuCLmaYXERERESl1JVndSEREREREikdBgoiIiIiItKAgQUREREREWlCQICIiIiIiLShIEBERERGRFhQkiIiIiIhICwoSRERERESkBQUJIiIiIiLSgoIEERERERFpQUGCiIiIiIi0YO5e7DJ0eWa2BVibh6wGA1vzkI+UBx3vrkfHvGvR8e5adLy7lmIe7zHuPqStRAoSOhEze8ndZxS7HNIxdLy7Hh3zrkXHu2vR8e5ayuF4q7qRiIiIiIi0oCBBRERERERaUJDQudxZ7AJIh9Lx7np0zLsWHe+uRce7ayn54602CSIiIiIi0oKeJIiIiIiISAtVxS6AiIhIV2dmlcCHgTOBPcCr7v5QcUslIl2ZniR0AmY22cweNLOlZrbEzOaZ2VnFLpfkzswqzewaM3vVzHab2Wozu8/MxqZIP9rM7jKzZWb2mpm9YGYXdWypJZ/M7FYzW5Ni3clm9oSZLTaz5Wb2mJkd08FFlDwwswoz+yKwEjgP+Dd3vz4xQNB5vnxZ5KNm9qSZPWVmL5rZc2b2mRAcJqbX+bwMmdnx4bx8Wpo0WR1bMxtoZj8J6ReY2SIzu8rMrDDfIoG7ayrjCZhFdNfp+0B3osDvaqARuKLY5dOU0zGtAH4TjuF6YB/gYdoDTE9IP4VoQJZfA70AAz4Str+p2N9HU06/gQvC8V6TZN1ngAbgGqAS6Ab8G7Af+HCxy64pq+M8CJgD7ATOSpNO5/kynoD/AlYDk+KWnQ7sBX6YkFbn8zKbgBPD33Hs//TMFOmyOrbAyPC7mQsMCMuODeeCXxLaFRf0uxV752pqx8GDfsB24CWgKm65Ab8H6oDxxS6npqyP68eBZ4hGRISoWuCFwLZwAlocl7YSWAa8BfRJyOcnIf1pxf5OmrI6/uOAZ5MFCcD4EAw8Ev8PIvxGXgLeBfoW+ztoyug49wzHuRE4N006nefLeAKOD3/LlyVZd0tYF7sA1Pm8zCZgNHBWmB9IFSTkcmyBPwO7gZEJy/9PSP+ZQn8/VTcqb5cAA4AH3L0+ttCjX9H9RBcO1xSnaJKL8AjxSuAf3X0tgLvXe1T14KqQ7EgzGxlenwtMBB51910J2d0f5l8tcLElT8ysO/AL4PMpklxDdCf51+HvHIh+I0RPn/oBlxa6nJIXNwMnAT9298fTpLsEnefL2clhvjPJuhVhfmiY63xeZtx9nbv/xd3XARvTJM3q2JrZNKLgY567b0hI/0AsfaGrHSlIKFPhhxH7x/CXJEnmhvllZtarY0oleTCM6CSyNcm6R4nuGgIcHOZfCvNkv4GXie4sn2dmY/JaSimUfwf+x90XJ64ws97AZUR3kOYl2fbPYf6lDquvKjkxswnAF8Lbn6VJp/N8+YtdEM5Osu59ROfoZeG9zuflLd2YAtke25R/9+7+DvAaMI3mILQgFCSUr6HAhPD6zcSV4SJzG9AXOKwDyyXt4O6b3P0/U6w7QPTosQFYES4gTgqrk/0GGogaQ1YAatRa4szsI8AI4K4USY4k+nve6u7bkqyPXWiMJ7rzLKXrMqInAMuILgweCo3QnzSz75jZQSGdzvPl77dEd5gvNLMfmVkVgJlNJAocPufuu3Q+77xyPLYnp0ofxM73x+WpmEkpSChfsWizjqjxUzI7EtJKGTOz4UB/4Al3fxcYQlSvGaK7EMnoN1AGzGwU8M/AVfHViBLEjmGqY70bqE9IK6XpQ2G+g6h9yUVEFxF/B74LvGxmw9B5vuyFc/VMYAPwNeBxMzuHqD3CB725FyudzzuvrI5tCCraOt93yG9BQUL5Gh3mO9JcVOiE0rnMJroIvCG8Hx23bkfr5C2W6zdQosKdxXuAL7p7snrLMU1/88lWhvOAjneJCxcA48Pbr7n7ao+85+7fBX4FHA7chM7znYK7LyPqzWgd8AHgcaI7wa/EJdP5vPPK9tgOojmoKOpvQUFC+ToQ5o1p0sTWqX5ymQv1jb8K/LO7LwqLD8QlSfU70G+g9N0EPO7uL7WRTn/zncMgoE94vSTJ+lgbhctpboOkY17++gMLiOqmHyDqoebPZtY3rNf5vPPK9tjWJVmeLn3BKEgoX+vCvH+aNLF6yWsLXBYpvJuAJ4EfxC1bF/c61e9Av4ESZmZnE7U1SNoOJUHav/lwh1rHu/Tti3vdLcn6V8O8EtgSXus8X8bM7ENETws/5+63AqcSdYV5GvCEmXVD5/POLKtjG54o78w0fTvLlpaChPIV+9FVm1l1ijQDEtJKGTKz2cAo4OqEKgc7gdrwuq0TiX4DpenbRA3VloSRdJumsH5k3Pu3w7JUx7onUfeooONdsty9luaL/2FJ1u8jaowMUeNm0Hm+bJnZ0cBjRE8LNwO4+3xgBtFgmacAn0Xn884sl2Pb1o3gDvktKEgoXztoflQ9NnGlmfUjeqy9C1iauF7Kg5nNBM4DPhnfRzo01UF/Jrwdl2TbirDcibpYk9JTQTSq5mFJJoguEmPvBxH93Q8xsz6ts2rqa301zReZUppWhvmoFOsrw/xNdJ4vd58kOp4vxi8MAUOsfdnZOp93Xjke27+lSh/Ezvfz81TMpBQklKnwo7s1vJ2VJMlMouN7d7hzJWXGzE4kGljtEnevS1jXzczGk/43cBQwGHjM3VcXtLCSE3ef6e6WbApJ1sYt+yNR96iVRI0gE50V5v+dppGrlIZ7wvwjiStC96f9iZ4cbUbn+XIXu846KMm6VxLe63zeeWV7bG9Lld7MBgLTgUU0Bx8FoSChvN0LbAcuDpFovIuJesK5tdVWUvLMbDpRA7dLw/gI8eu6Ex3XHsAcojuI5yepjnBxmP+kwMWVjvNTokZwLQZmCn//FwHvEY3YLKXtHqIuMS80s8QxLc4J8++GYE/n+fIWuyP8oSTrxiWk0fm8vMXaGCVrTJzVsXX314naIc4ys8EJ6S8Mn/Hjgt8QcndNZTwB7ydqCHcDUdWECuBSopbvVxS7fJpyOqaTga1E1UaWJkxvEvWX/nRc+ilEFxF3ANXh5HE20cXDTcX+Pppy/h04sCbJ8s8SDah3Sfh7rwK+A+wHPlzscmvK+PieEf6W5wIDwrLh4W/8YaAyLq3O82U6hWP1u/D3fCVgYflo4HWisTG6x6XX+bwMJ6KAryEc58tTpMnq2BJVRV0LPAH0CcumEVUv/GXst1TIKfZjlTJmZlOBG4mGeK8nGt3x/7r73KIWTLJmZocAL5GkQWOCT7n7fXHbjSPqAelEYE+Y/sObB+qRMmNmTlTdaGySdacTXTCOILpoXEp05zmx+oKUMDM7gmjwtNOAFUQD4j0C3OEJ/5x1ni9fYSyUzwJXEI2OvYmok4HfAz/1qLF6fHqdz8uEmY0hGlX7KKIL/5gNwL+7+20J6bM6tuEpwo1E42vsIgo6fwrcmXiOKAQFCSIiIiIi0oLaJIiIiIiISAsKEkREREREpAUFCSIiIiIi0oKCBBERERERaUFBgoiIiIiItKAgQUREREREWlCQICIiIiIiLShIEBERERGRFhQkiIiIiIhICwoSRERERESkBQUJIiIiOTCz7mb2fjPrU+yyiIjkm4IEEZEchYvEb5jZAjN7xcyWmVmjmXmYPheXttLMnjSz5WY2upjlzpaZ9TGz/zKzX5nZVjN7zsyuKHa5oDj71cyOM7MHga3Ak8CgjvhcEZGOpCBBRCQHZmbAg8C/AJe7+3R3nwQcBcwPyUbEbVIFTARGAf06sqztYWbdgGeBA+7+aeBsYCrw3Qy3L/T/mWLs15eB7wC985FZB+wjEZGs6cQkIpKbc4Dzgfvc/eXYQnd/HfgQsBoYGbd8PzAJGOLur3VwWdvjAqLAZx6Au78S3p/a1oZm1hP4YSELV4z96u4N7r4UWN/evDpiH6X43N+Hp13/ZGa9zOxfzOx1M9tjZm+Z2Y/NrFdc+pPN7HdmttHM9pnZfDM7s6PLLSIdp6rYBRARKVOnhHl14gp3325m3wIuSVheB9QVvmh5FbsQ3B1b4O4r29ooPGm5FRhfoHI1KeJ+bWzPxh25j5KYGua1wMJQhl1Ad+AQ4KtET8Jmm9lNwD8DDcBeoAdwLPCEmR3m7us6uOwi0gH0JEFEJDcNYf5xM5ueZP0fSHEjpsyql4xoO0lS1wGX5bMgbSmz/QpF2EcAZtYXGBve3gSsA6a5e1+gL/DrsO4iM/s+8A3gS0D/kOZk4D2iAPnqDiy6iHSgcjuhioiUirlh3gN41syuCXeGAXD3A+7+wdh7MxttZt8yswXAaXHLx4ZqH6vM7GkzqwnT6rC83swmxqWvNLOvmdltZvaCma0MrwdmWnCLfM7MHjOzp8xshZk9YWYXxKX5gZnV0PzE5D/jynZYG/n/A/Cx8HZa3HbjwvpeZnalmS0K7y8PDaLnxPahmX3AzP5kZs+EajAvmtm5ST6r1X41syPN7JuhQfO9ZtbTzG42szfNbJuZ3WlmB2W6v0Kenwn5PWtmf43fV0nStln2DPZRRt8/R1OA2G/1NeBsd18I4O57gP8T1lUA3wQ+6u63ufvukObvwN0hzaQ8lUlESo27a9KkSZOmHCaiCyWPm+YAo5KkOwH4BdHTBwdmxq0bCzwEWNyyauDNkPZf45YfBDwKHBW37PKQ7gWgKoMyVxP1yPMiMDws6wncFfL5NVARl74mscwZ7puZYbuahOUfJ7pzHdtnpxPdlY69PwT4VHh9edimG/CnsOyItvYrcARwT1j2CvAIMDuk/9/E/drG9zDgdmAjcGTc/nqAqLqRA2Pj0mdU9jb2UcZ55Pi7/WLI6wAwJkWaPSHN7SnWfyus/0Ox/w41adJUmElPEkREcncF8AOiiyWADwBLzOzrZlYZS+Tuz7v7ZcDrSfJoAB5wd49bdgPRHdolwPfiln8DWOotG+jeDWwBjgP+MYMyXwe8H7jC3TeF8u0FrgSeBz4ZXheEu98PnBW36CKiwOBDwLXABuCasG5B2KaOKDiKpY/llXS/uvsS4LfhbS/gs+7+gLs/D3wlLJ+dYZFnA58HvupRo/TY/vo8zcc9XkZlb0M+8kgn1h7hz+6+NnGlmQ0mCoQA/l+KPMaEudojiHRSargsIpIjd28ArjWzh4nuNk8juij9D+BEM/uUu++L22R7kjzeAt6KvTezqTTfpb3Mo957Yr3gXAusN7PjE7LZQXTndyrNF8ethCo21wEbPVQviStHo5n9N9Hd9u+a2R0JgUs+rYh7fbO7v0f0FGZOKOevgLdpefG/J8yTtZFotV9pbmi90d13xi1fFeZjaENo4/D98NkPx69z951mtoGo69V42ZY9mXzkkU4sSHi8jfWb3f3FFGmOCvPFeSiPiJQgBQkiIu3k7i+Y2XHA14nGD+hB1HXodrK4K29mVUR3bquAn4Q73zETiaob3eXuuXaZeSRR1ZXNKdbXhPkwYDhRFZu8c3ePNd9w91bdiLr7bcBtAGbWG/gEcGFY3ao3qSw/uz58dib//8YSBQFL3P1AkvWtejfKR9kL+f1D4DM5vH0lRbJYkPBqmqymhPmC9pRHREqXggQRkTwIVUJuNrNHgGeAwcDnzOxmd1+VfusmXwGOIbrbfUPCulgD0XHtKObYMB+WYv0moq5EuxHdsS5IkJAJMxtPVO2miuiC+QBRNamOdGSY78pmo3yUvYDffwJRsNkILEqRZlqYJw0AzGws0cB1jUQNn0WkE1KbBBGRLJnZ8Wb2uWTrPBpk6wtxi6YlS5ckz/FE3VFC1GB1d0KS98J8CimY2XgzS3enOTa+wbBQfSmREwUJDixtu9SFEXr+WQgsdvcvhX1aDLHqVhlX8clH2Qv8/WNPCZZ71JNRMmmDhLj1K5L8TkWkk1CQICKSvbeBm9J0o/mXuNdtNuwM3X7eSdRY9A53r0lYP5nmevwnm9kZSfLoSxRk7E/zUW8QDYZlwEeSrB9IdJd5QR4v/qztJHGJoy5AfwdsdfdUjWY7SuzifISZDWkrcTvK3rSPOuD7x4KEVE8JugOHp0sTl8fCFOtFpBNQkCAikr0NRNUtbo/1658gViVoLS0vpLqFeWJVz8uIRjZeT9RouUm4+J8NrAaeC4t/ZWZnxY0pMIVo3IZfpWtsHBpR3xjefjG+B6Yg1vf/1+KWdU+YZypWRafVxXX8PrPWA6AdC1SGZPH7dnCaciTbr0n/v5lZt2TLU1gFzA95tWhbEsoW+7xYmbIte7J9lPX3N7NBZtajzW8TSRsk0NxuZQ+wLMc8RKQTUJAgIpIld28ElgOfBn5tZsNj68xsNFFPR7uBj4W2CrE7tLHgYUJc+uFEvSEBfD709BNbV0XUEHpzuPj/ash3JPBnYLuZbSIKRH7r7nMyKP6PicYNOA34UazaUQg0/gW4PvYkI6yLlXVq66zSWk701GKimfUzsx5mNiusOyIu3eSE7RYTVfMZDfzQzM40s38HTgrrjwkDjR0byph0v8a9HpSQ//tiL8wssWeiFsJxvgaoBW4ws/PiLtw/Q3M1pIvMbFi2ZSfJPiJqMJ7N97+UqAvctdbGIHdBWxf4sapEr4Xvny4PPUkQ6cyKPVCDJk2aNJXjRNTVaGwAsHqiMQ1eJrpwuhMYGZf2OqKnD7H0dUR3/ofSPOjXfqLehWLTfOCdsO6iuLwmEXXH+RZR16fzgFlZlt2AS4gGVVsF/DF8n1Pi0vxX3OfHpteIumXN9HOuArYSNeS+jqgq03/TPAiZE10kvwB0j9vuivD9thAFNNOJeox6JnznXwB90+zXi4mqVsWW30XUO9QJRFXBYsvnAcdn8D0mEA14t56ose8vgHPC67nAv9E8kFtGZW9jH2WcB1GvR9vC9/lJG99jQNx3H54izX+SfhC1PnHHb2Rb+06TJk3lO5l7obrBFhER6bzMzLxE/oma2ZXACHf/brHLIiKdg6obiYiI5KCEAoQeRE8UflHssohI56EgQUREpEyZWS/gPuCH7t5mT1oiIplSdSMREZEyFRpuD3T3TcUui4h0LgoSRERERESkBVU3EhERERGRFhQkiIiIiIhICwoSRERERESkBQUJIiIiIiLSgoIEERERERFpQUGCiIiIiIi08P8BH2dONRZOp6cAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "m = [i for i in range(2, 101)]\n", "fig, ax = plt.subplots(figsize=(12, 8))\n", "plt.xlabel('Size of train data, $m$')\n", "plt.ylabel('Error rate, $Q$')\n", "plt.errorbar(x = m, y = error_mean, yerr = np.array(error_std))\n", "ax.set_title('Зависимость ошибки от объема выборки')\n", "plt.grid()\n", "plt.show()\n", "fig.savefig('ModelBreadSw.png')" ] } ], "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": "3.6.5" } }, "nbformat": 4, "nbformat_minor": 2 }