{"cells": [{"cell_type": "markdown", "metadata": {}, "source": ["# Pr\u00e9diction d'une dur\u00e9e\n", "\n", "Ce notebook cherche \u00e0 pr\u00e9dire la dur\u00e9e de stockage de paquets pr\u00e9par\u00e9s par un magasin. Chaque paquet met plus ou moins de temps \u00e0 \u00eatre pr\u00e9par\u00e9. Si la commande arrive le soir et ne peut \u00eatre finie avant la fin de la journ\u00e9e, elle est report\u00e9e sur la journ\u00e9e suivante. C'est la particularit\u00e9 de ce jeu de donn\u00e9es."]}, {"cell_type": "markdown", "metadata": {}, "source": ["* L'heure de la commande est uniform\u00e9ment r\u00e9partie sur la semaine et la journ\u00e9e.\n", "* La dur\u00e9e suit une loi $\\Gamma(3, 0.5)$."]}, {"cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [{"data": {"text/html": ["
\n", ""], "text/plain": [""]}, "execution_count": 2, "metadata": {}, "output_type": "execute_result"}], "source": ["from jyquickhelper import add_notebook_menu\n", "add_notebook_menu()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Distribution gamma\n", "\n", "La loi [Gamma](https://fr.wikipedia.org/wiki/Loi_Gamma) est souvent utilis\u00e9e pour mod\u00e9liser une dur\u00e9e de vie."]}, {"cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": ["%matplotlib inline"]}, {"cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["from numpy.random import gamma\n", "g = gamma(6, 0.25, 1000)\n", "\n", "import matplotlib.pyplot as plt\n", "import pandas\n", "fig, ax = plt.subplots(1, 1, figsize=(8, 3))\n", "df = pandas.DataFrame(dict(g=g))\n", "df.hist('g', bins=50, ax=ax)\n", "ax.set_title('Distribution gamma(6, 0.25)')\n", "ax.set_xlim([0, 4]);"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On simule un jeu de donn\u00e9es avec la fonction [duration_selling](http://www.xavierdupre.fr/app/papierstat/helpsphinx/papierstat/datasets/duration.html?papierstat.datasets.duration.duration_selling)."]}, {"cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [{"data": {"text/html": ["\n", "\n", "
\n", " \n", " \n", " \n", " commande \n", " reception \n", " true_duration \n", " \n", " \n", " \n", " \n", " 0 \n", " 2018-02-05 13:15:09.062249 \n", " 2018-02-05 14:18:58.540886 \n", " 1.063744 \n", " \n", " \n", " 1 \n", " 2018-02-05 17:16:52.648499 \n", " 2018-02-05 18:21:28.592673 \n", " 1.076651 \n", " \n", " \n", " 2 \n", " 2018-02-05 09:37:59.627662 \n", " 2018-02-05 10:58:42.387613 \n", " 1.345211 \n", " \n", " \n", " 3 \n", " 2018-02-05 09:02:20.357816 \n", " 2018-02-05 10:03:05.420104 \n", " 1.012517 \n", " \n", " \n", " 4 \n", " 2018-02-05 13:52:33.335866 \n", " 2018-02-05 15:20:27.454535 \n", " 1.465033 \n", " \n", " \n", "
\n", "
"], "text/plain": [" commande reception true_duration\n", "0 2018-02-05 13:15:09.062249 2018-02-05 14:18:58.540886 1.063744\n", "1 2018-02-05 17:16:52.648499 2018-02-05 18:21:28.592673 1.076651\n", "2 2018-02-05 09:37:59.627662 2018-02-05 10:58:42.387613 1.345211\n", "3 2018-02-05 09:02:20.357816 2018-02-05 10:03:05.420104 1.012517\n", "4 2018-02-05 13:52:33.335866 2018-02-05 15:20:27.454535 1.465033"]}, "execution_count": 5, "metadata": {}, "output_type": "execute_result"}], "source": ["from papierstat.datasets import duration_selling\n", "df = duration_selling()\n", "df.head()"]}, {"cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [{"data": {"text/plain": ["(2481, 3)"]}, "execution_count": 6, "metadata": {}, "output_type": "execute_result"}], "source": ["df.shape"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Quelques statistiques\n", "\n"]}, {"cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": ["stat = df.copy()\n", "stat[\"wk\"] = stat.commande.dt.weekday\n", "stat[\"wk2\"] = stat.reception.dt.weekday\n", "stat[\"hr\"] = stat.commande.dt.hour\n", "stat[\"hr2\"] = stat.reception.dt.hour"]}, {"cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [{"data": {"text/html": ["\n", "\n", "
\n", " \n", " \n", " \n", " jour \n", " commande \n", " reception \n", " \n", " \n", " \n", " \n", " 0 \n", " 0 \n", " 502 \n", " 498 \n", " \n", " \n", " 1 \n", " 1 \n", " 531 \n", " 529 \n", " \n", " \n", " 2 \n", " 2 \n", " 535 \n", " 544 \n", " \n", " \n", " 3 \n", " 3 \n", " 443 \n", " 458 \n", " \n", " \n", " 4 \n", " 4 \n", " 470 \n", " 452 \n", " \n", " \n", "
\n", "
"], "text/plain": [" jour commande reception\n", "0 0 502 498\n", "1 1 531 529\n", "2 2 535 544\n", "3 3 443 458\n", "4 4 470 452"]}, "execution_count": 8, "metadata": {}, "output_type": "execute_result"}], "source": ["wk = stat.groupby('wk').count()[[\"commande\"]]\n", "wk2 = stat.groupby('wk2').count()[[\"reception\"]]\n", "wks = wk.merge(wk2, left_on=wk.index, right_on=wk2.index)\n", "wks.columns = ['jour'] + list(wks.columns)[1:]\n", "wks"]}, {"cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["import matplotlib.pyplot as plt\n", "fig, ax = plt.subplots(1, 1, figsize=(8, 3))\n", "wks.plot(x=\"jour\", y=[\"commande\", \"reception\"], kind=\"bar\", ax=ax)\n", "ax.set_xlabel('jour de la semaine')\n", "ax.set_title(\"Distribution des commandes sur la semaine\")\n", "ax.set_ylim([440,560]);"]}, {"cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [{"data": {"text/html": ["\n", "\n", "
\n", " \n", " \n", " \n", " heure \n", " commande \n", " reception \n", " \n", " \n", " \n", " \n", " 0 \n", " 9 \n", " 260 \n", " 249 \n", " \n", " \n", " 1 \n", " 10 \n", " 233 \n", " 262 \n", " \n", " \n", " 2 \n", " 11 \n", " 244 \n", " 264 \n", " \n", " \n", " 3 \n", " 12 \n", " 239 \n", " 223 \n", " \n", " \n", " 4 \n", " 13 \n", " 268 \n", " 237 \n", " \n", " \n", " 5 \n", " 14 \n", " 249 \n", " 266 \n", " \n", " \n", " 6 \n", " 15 \n", " 249 \n", " 241 \n", " \n", " \n", " 7 \n", " 16 \n", " 247 \n", " 256 \n", " \n", " \n", " 8 \n", " 17 \n", " 228 \n", " 262 \n", " \n", " \n", " 9 \n", " 18 \n", " 264 \n", " 221 \n", " \n", " \n", "
\n", "
"], "text/plain": [" heure commande reception\n", "0 9 260 249\n", "1 10 233 262\n", "2 11 244 264\n", "3 12 239 223\n", "4 13 268 237\n", "5 14 249 266\n", "6 15 249 241\n", "7 16 247 256\n", "8 17 228 262\n", "9 18 264 221"]}, "execution_count": 10, "metadata": {}, "output_type": "execute_result"}], "source": ["hr = stat.groupby('hr').count()[[\"commande\"]]\n", "hr2 = stat.groupby('hr2').count()[[\"reception\"]]\n", "hrs = hr.merge(hr2, left_on=hr.index, right_on=hr2.index)\n", "hrs.columns = ['heure'] + list(wks.columns)[1:]\n", "hrs"]}, {"cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["fig, ax = plt.subplots(1, 1, figsize=(8, 3))\n", "hrs.plot(x=\"heure\", y=[\"commande\", \"reception\"], kind=\"bar\", ax=ax)\n", "ax.set_xlabel('heure du jour')\n", "ax.set_title(\"Heures des commandes sur la semaine\")\n", "ax.set_ylim([200,300]);"]}, {"cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["fig, ax = plt.subplots(1, 3, figsize=(12, 3))\n", "stat[\"duree\"] = stat.reception - stat.commande\n", "stat[\"hours\"] = stat.duree.dt.total_seconds() / 3600\n", "stat['hours'].hist(bins=500, ax=ax[0])\n", "stat[stat['hours'] < 10]['hours'].hist(bins=50, ax=ax[1])\n", "stat['true_duration'].hist(bins=50, ax=ax[2])\n", "ax[0].set_title(\"Distribution de la dur\u00e9e\")\n", "ax[1].set_title(\"Distribution de la dur\u00e9e\\ntronqu\u00e9e \u00e0 10h\")\n", "ax[2].set_title(\"Distribution de la dur\u00e9e\\nayant servi \u00e0 l'estimation\")\n", "ax[0].set_xlabel(\"Heures\")\n", "ax[1].set_xlabel(\"Heures\")\n", "ax[2].set_xlabel(\"Heures\");"]}, {"cell_type": "markdown", "metadata": {}, "source": ["La distribution montre trois pics : le premier correspond aux paquets r\u00e9ceptionn\u00e9s dans la journ\u00e9e, le second le lendemain et le troisi\u00e8me le week-end. La dur\u00e9e de livraison d\u00e9pend uniquement de l'heure d'arriv\u00e9e de la commande. Si elle arrive assez t\u00f4t, la livraison est faite al\u00e9atoirement dans la journ\u00e9e, sinon, on passe au jour suivant."]}, {"cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["import seaborn\n", "fig, ax = plt.subplots(1, 2, figsize=(12, 4))\n", "seaborn.boxplot(x=\"hr\", y=\"hours\", data=stat, palette=\"muted\", ax=ax[0])\n", "seaborn.boxplot(x=\"hr\", y=\"hours\", data=stat, palette=\"muted\", ax=ax[1])\n", "ax[1].set_ylim([0, 15])\n", "ax[0].set_title(\"Distribution de la dur\u00e9e de livraison\\nen fonction de l'heure d'arriv\u00e9e\")\n", "ax[1].set_title(\"Distribution tronqu\u00e9e de la dur\u00e9e de livraison\\nen fonction de l'heure d'arriv\u00e9e\");"]}, {"cell_type": "markdown", "metadata": {}, "source": ["## Pr\u00e9diction de la dur\u00e9e\n", "\n", "Il n'y a pas grand-chose \u00e0 pr\u00e9dire puisque tout est al\u00e9atoire except\u00e9 le fait que le fait qu'un paquet soit r\u00e9ceptionn\u00e9 le lendemain ou le lundi suivant. On construit donc une variable correspondant \u00e0 l'heure de commande et le jour de la semaine. On compare un mod\u00e8le lin\u00e9aire et un arbre de d\u00e9cision."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Comparaison d'un mod\u00e8le lin\u00e9aire et d'un arbre de pr\u00e9cision"]}, {"cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [{"data": {"text/html": ["\n", "\n", "
\n", " \n", " \n", " \n", " commande \n", " reception \n", " true_duration \n", " heure \n", " wk \n", " duree \n", " \n", " \n", " \n", " \n", " 0 \n", " 2018-02-05 13:15:09.062249 \n", " 2018-02-05 14:18:58.540886 \n", " 1.063744 \n", " 13.250000 \n", " 0 \n", " 1.063744 \n", " \n", " \n", " 1 \n", " 2018-02-05 17:16:52.648499 \n", " 2018-02-05 18:21:28.592673 \n", " 1.076651 \n", " 17.266667 \n", " 0 \n", " 1.076651 \n", " \n", " \n", " 2 \n", " 2018-02-05 09:37:59.627662 \n", " 2018-02-05 10:58:42.387613 \n", " 1.345211 \n", " 9.616667 \n", " 0 \n", " 1.345211 \n", " \n", " \n", " 3 \n", " 2018-02-05 09:02:20.357816 \n", " 2018-02-05 10:03:05.420104 \n", " 1.012517 \n", " 9.033333 \n", " 0 \n", " 1.012517 \n", " \n", " \n", " 4 \n", " 2018-02-05 13:52:33.335866 \n", " 2018-02-05 15:20:27.454535 \n", " 1.465033 \n", " 13.866667 \n", " 0 \n", " 1.465033 \n", " \n", " \n", "
\n", "
"], "text/plain": [" commande reception true_duration \\\n", "0 2018-02-05 13:15:09.062249 2018-02-05 14:18:58.540886 1.063744 \n", "1 2018-02-05 17:16:52.648499 2018-02-05 18:21:28.592673 1.076651 \n", "2 2018-02-05 09:37:59.627662 2018-02-05 10:58:42.387613 1.345211 \n", "3 2018-02-05 09:02:20.357816 2018-02-05 10:03:05.420104 1.012517 \n", "4 2018-02-05 13:52:33.335866 2018-02-05 15:20:27.454535 1.465033 \n", "\n", " heure wk duree \n", "0 13.250000 0 1.063744 \n", "1 17.266667 0 1.076651 \n", "2 9.616667 0 1.345211 \n", "3 9.033333 0 1.012517 \n", "4 13.866667 0 1.465033 "]}, "execution_count": 14, "metadata": {}, "output_type": "execute_result"}], "source": ["data = df.copy()\n", "data[\"heure\"] = data.commande.dt.hour + data.commande.dt.minute / 60\n", "data[\"wk\"] = data.commande.dt.weekday\n", "data['duree'] = (data.reception - data.commande).dt.total_seconds() / 3600\n", "data.head()"]}, {"cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [], "source": ["X = data[[\"heure\", 'wk']]\n", "y = data[\"duree\"]"]}, {"cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": ["from sklearn.model_selection import train_test_split\n", "X_train, X_test, y_train, y_test = train_test_split(X, y)"]}, {"cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [{"data": {"text/plain": ["0.20497865247039226"]}, "execution_count": 17, "metadata": {}, "output_type": "execute_result"}], "source": ["from sklearn.linear_model import LinearRegression\n", "clr = LinearRegression()\n", "clr.fit(X_train, y_train)\n", "clr.score(X_test, y_test)"]}, {"cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [{"data": {"text/plain": ["0.6591302839171569"]}, "execution_count": 18, "metadata": {}, "output_type": "execute_result"}], "source": ["from sklearn.tree import DecisionTreeRegressor\n", "clr2 = DecisionTreeRegressor(max_depth=3)\n", "clr2.fit(X_train, y_train)\n", "clr2.score(X_test, y_test)"]}, {"cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [{"data": {"image/png": "iVBORw0KGgoAAAANSUhEUgAAAtoAAAEKCAYAAAAsOPKBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xl4lOW9//H3PZOEkEAgCUtCQgjIvgQIhMWQUXBBe1BRERWrUheq1qpVT7X2J7VWe2x7qlZrD7VqpRY3tIh6jlZRcBIWIWERWWRNYAIECAECISSZuX9/TIjsDJDkScLndV1ek2fmmXm+YybDZ+65n+9trLWIiIiIiEjtcjldgIiIiIhIU6SgLSIiIiJSBxS0RURERETqgIK2iIiIiEgdUNAWEREREakDCtoiIiIiInVAQVtEREREpA4oaIuIiIiI1AEFbRERERGROhDmdAG1pU2bNjY1NdXpMkRERESkCcvLy9tprW0byr5NJminpqaSm5vrdBkiIiIi0oQZYwpC3VdTR0RERERE6oCCtoiIiIhIHVDQFhERERGpA01mjraIyOmqrKzE5/NRXl7udCnigMjISJKTkwkPD3e6FBFpohS0ReSc5fP5aNmyJampqRhjnC5H6pG1luLiYnw+H507d3a6HBFpohydOmKM6WGMWXrYf3uNMQ8YY+KMMZ8bY9ZWX8Y6WaeINE3l5eXEx8crZJ+DjDHEx8fr2wwRqVOOBm1r7XfW2gHW2gHAIKAMmAE8Cnxhre0GfFG9LSJS6xSyz1363YtIXWtIJ0NeBKy31hYAVwFTq6+fCox1rCoRkQZoxYoVfPTRR06XISJS7wr2FnCg6oDTZYSkIQXtG4C3qn9ub63dClB92e54dzDGTDLG5Bpjcnfs2FFPZYqI1J4WLVoAsGXLFsaNGxfSfTZt2sTTTz/NBRdccMp9J0+ezKxZs86qRhERJ1X6K5m/ZT6/W/g7xswYw5gZY5hXOM/pskJirLVO14AxJgLYAvSx1hYZY3Zba1sfdnuJtfak87QHDx5stTKkiJyOVatW0atXL0draNGiBfv27av34/r9ftxud70ft6FpCK8BETmWP+DH7XKzdd9Wxs4cS1lVGRGuCDISM8hKyuKSTpfQLuq447B1zhiTZ60dHMq+DWVE+3JgsbW2qHq7yBiTCFB9ud2xykREDpNXUMJLs9eRV1BSq4+bn59P3759AXj99de55ppruOyyy+jWrRs///nPa/b77LPPGD58OOnp6Vx33XU1If3JJ58kIyODvn37MmnSJA4NokycOJH33nsPgNTUVJ588klGjBjB9OnTWb9+PZdddhmDBg0iKyuL1atX1+pzEhEJlT/gZ9mOZby45EXGfzSeX8//NQAJ0Qlc3+N6Xhz1Itk3ZDPl4inc1Osmx0L26Woo7f1u5PtpIwAfArcCz1RfznSiKBGRw+UVlHDTKwuoqAoQEeZi2h3DGNSpbpoiLV26lCVLltCsWTN69OjBT3/6U5o3b85TTz3FrFmziI6O5ne/+x3PPvsskydP5t5772Xy5MkA3HzzzXz88cdcccUVxzxuZGQkOTk5AFx00UVMmTKFbt268fXXX3PPPffw5Zdf1snzERE5kefynmPG2hmUHCzBZVwMaDuAvm2CAw/GGB4c/KDDFZ45x4O2MSYKuAT48WFXPwO8a4y5HdgEXOdEbSIih1uwoZiKqgABC5VVARZsKK6zoH3RRRfRqlUrAHr37k1BQQG7d+9m5cqVZGZmAlBRUcHw4cMBmD17Nr///e8pKytj165d9OnT57hB+/rrrwdg3759zJs3j+uu+/7t9eDBg3XyXEREINi/fu3utWT7slm6fSl/GvUnXMaF27jJTMrEk+zh/A7n06pZK6dLrTWOB21rbRkQf9R1xQS7kIiINBjDusQTEeaisipAeJiLYV3iT32nM9SsWbOan91uN1VVVVhrueSSS3jrrbeO2Le8vJx77rmH3NxcOnbsyBNPPHHC/tDR0dEABAIBWrduzdKlS+vsOYiIAKwpWcM7q98huzCbrfu3AtAzrifFB4ppG9WW+9Lvc7jCutNQ5miLiDR4gzrFMu2OYTx4aY86nTZyIsOGDWPu3LmsW7cOgLKyMtasWVMTqtu0acO+fftq5mSfTExMDJ07d2b69OlAcKRp2bJldVe8iJwzfKU+3lr9FutKgu9VRfuL+GjDR/SK68UTw59g1rhZTL9iOm2j2jpcad1zfERbRKQxGdQptt4D9iFt27bl9ddf58Ybb6yZ5vHUU0/RvXt37rzzTvr160dqaioZGRkhPd60adO4++67eeqpp6isrOSGG26gf//+dfkURKQJ8gf8LN6+GK/Pi9fnZcOeDQA8NOghusZ2ZXiH4eTckEOEO8LhSutfg2jvVxvU3k9ETpdau4leAyJnZueBnewo20Gv+F5U+CsY8fYIKgOVDG4/GE+yB0+yh04xnZwus06cTns/jWiLiIiIyEkFbIBVxatqRq2/Lf6WHrE9eO/K94hwR/C3S/9G19ZdiQ6PdrrUBkVBW0RERESOUVZZRlR4FACPz32cD9d/iMHQr20/7h1wL55kT82+/dtq2tnxKGiLiIiICNZaNu7dSLYvG6/Py+Lti/m/q/+PxBaJXHXeVQxLHMaIpBHERjpznkpjpKAtIiIico77Zsc3POJ9BN8+HwBdW3fllt634DLBBnVDEoc4WV6jpaAtIiIicg7Ztn9bzVzrUSmjuKbbNXRo0YHzWp/Hj/r+iKykLBJbJDpdZpOgoC0iIiLSxFlreXHJi8zxzWFtyVoAklok1cyzbtO8DX++6M9OltgkKWiLiDQCqamp5Obm0qZNG6dLEZFGoKS8hJzCHIrKirij3x0YY8gryqN1s9Y8NOghPMkeOrfqjDHG6VKbNAVtEZEGzu/3h7RfVVUVYWF6Wxc5V23Ys4FZBbPw+rx8s+MbLJaE6AQm9plImCuM10a/htvldrrMc4rekevC5oWQnw2pWdBxyLHbItJ41cHf89ixY9m8eTPl5eXcf//9TJo0iRYtWvDggw/y73//mz/+8Y8A/OEPf2D27NkAvPnmm3Tt2pWJEycSFxfHkiVLSE9P58knn+SnP/0py5cvp6qqiieeeIKrrrqqVuoUkYalbKOX+Uv+yvDwWKISBzNrwwe8eGA9fd0x3JV6BZ4+N9I7vjeuvH/Akn/gbpkA8d1g2zfQ6yoYPNHpp9DkKWjXts0LYeqV4K8AdwRc9gx8+uj327d+qLAt0lgd/fddS3/Pr732GnFxcRw4cICMjAyuvfZa9u/fT9++fXnyySdr9ouJiWHhwoX84x//4IEHHuDjjz8GYM2aNcyaNQu3281jjz3GqFGjeO2119i9ezdDhgzh4osvJjpai0iINAUFewvw+rxkr/uY3F0rqDSG54t2cFHeG1zncnEN0CYQgI1roMtVkPcP+Pj+Yx9o/ZfBS4XtOuVyuoAmJz87+I+w9QcvV808cjs/2+kKReRMHf33XUt/zy+88AL9+/dn2LBhbN68mbVr1+J2u7n22muP2O/GG2+suZw/f37N9ddddx1ud/Dr4M8++4xnnnmGAQMGcOGFF1JeXs6mTZtqpU4RqX8V/gp2le8CYMPuDYyZMYbfL/o9RfsKmbC3lFe3FuEpOwBAbCAQDNkA/srge9SqmSd+8JPdJrVCI9q1LTUrONJ1aMSr11VQMP/77dQspysUkTN19N93Lfw9z5kzh1mzZjF//nyioqJqwnFkZGRNeD7k8JOWDv/58NFqay3vv/8+PXr0OOvaRMQZ28u21ywas2DrAi5NvZTfZP6Gzq0686vhv2Jo4lA67t4Kr48B/8HjP4g7PPge1Tz++9Hro/XStLK6pqBd2zoOCX6dfPgczva9NUdbpCk43t/3WdqzZw+xsbFERUWxevVqFixYcMJ933nnHR599FHeeecdhg8fftx9Ro8ezYsvvsiLL76IMYYlS5YwcODAs65TROqOtbbmw/N9X97H7M3BczHaR7XnP7r8B5d0ugQIfsAe131c8E4tO8LEj2HZW4CFhAGwbRns2w4t2kL/CcH3qEPvU0v+AZqjXe8cD9rGmNbAK0BfwAK3Ad8B7wCpQD4w3lpb4lCJp+/wF/bxtkWk8arlv+fLLruMKVOmkJaWRo8ePRg2bNgJ9z148CBDhw4lEAjw1ltvHXefxx9/nAceeIC0tDSstaSmptbM5RaRhmPPwT3M2zIPr8/L6l2ref/K93EZF0MShpDWNg1PsodurbudvP1eqO9HgycqVDvEWGudLcCYqUC2tfYVY0wEEAU8Buyy1j5jjHkUiLXWPnKyxxk8eLDNzc2th4pFpKlYtWoVvXr1croMcZBeA1Lf5m+Zz5RlU1i2Yxl+66d1s9aMSBrBL4b+gpiIGKfLkxAYY/KstYND2dfREW1jTAzgASYCWGsrgApjzFXAhdW7TQXmACcN2iIiIiINyYGqAyzcuhCvz8u13a+ld3xvAjbAgaoD3Nb3NjzJHvq16afe1k2Y01NHugA7gL8bY/oDecD9QHtr7VYAa+1WY0w7B2sUERERCcmBqgN8sO4DvD4vi7Yt4qD/IM3DmjOw/UB6x/cmMymTzKRMp8uUeuJ00A4D0oGfWmu/Nsb8CXg01DsbYyYBkwBSUlLqpkIRERGRE6gMVLJ0+1LKq8rJSs7Cbdw8l/cc7aLacV3368hKzmJw+8FEuCOcLlUc4HTQ9gE+a+3X1dvvEQzaRcaYxOrR7ERg+/HubK19GXgZgnO066NgERERObftPLCTnMIcvD4v87fMZ1/lPnrH9yYrOYsIdwT/d83/0aZ5G6fLlAbA0aBtrd1mjNlsjOlhrf0OuAhYWf3frcAz1ZfqqC4iIiKOCNgAa0vW0iMu2J/+6QVPM2vTLNo1b8fo1NFkJWUxrMP3HYMUsuUQp0e0AX4KTKvuOLIB+BHBFSvfNcbcDmwCrnOwPhGRBqO0tJR//vOf3HXXXSdv+yUiZ6W0opT5W+bj9XnJKcyhuLyYz679jMQWify4/4+ZlDaJnnE99XcoJ+X4EuzW2qXW2sHW2jRr7VhrbYm1tthae5G1tlv15S6n6xQRcVpFRQX33HMPF1xwwQn/cW/RogUAW7ZsYdy4cSd9vOeff56ysrKa7R/84Afs3r279go+Cy+88AK9evXipptuOuE+ubm53HfffUBwhc158+bVV3nSBFlrqQpUAeD1efG87eGhrx7iy81fMiRhCL8d8VtimgXb7/WM60mv+F4K2XJKjvfRri3qoy0ip6ux9FD2+/3HLMd+Ii1atGDfvn0h7Zuamkpubi5t2jScr7kPPdeePXvyySef0Llz55Du98QTT9CiRQsefvjh0zpeY3kNSN046D/Iom2L8Pq8eH1ebu1zKzf2vJGdB3byz5X/JCs5i/5t+xPmaggTAKShOJ0+2o6PaIuInKvy8/Pp2bMnt956K2lpaYwbN65mhDk1NZUnn3ySESNGMH36dNavX89ll13GoEGDyMrKYvXq1QBs3LiR4cOHk5GRweOPP37EY/ft2xcIhteHH36Yfv36kZaWxosvvsgLL7zAli1bGDlyJCNHjqw55s6dOwF49tln6du3L3379uX555+vecxevXpx55130qdPHy699FIOHDhwzPOaOHEid911F1lZWXTv3r1mZcr8/HyysrJIT08nPT29ZgR6zpw5jBw5kgkTJtCvXz/uuusuNmzYwJVXXslzzz3HwoULOf/88xk4cCDnn38+3333Xc39xowZQ35+PlOmTOG5555jwIABZGdns2PHDq699loyMjLIyMhg7ty5tf77k8bLH/Bz35f3kfV2FnfPupsZa2fQtXVXklskA8E51g8MeoBB7QcpZMtZ0atHRKTajz790THXjU4dzQ09b+BA1QHumXXPMbdf1fUqxnYdS0l5CQ/OefCI2/5+2d9PeczvvvuOV199lczMTG677Tb+8pe/1IzKRkZGkpOTA8BFF13ElClT6NatG19//TX33HMPX375Jffffz933303t9xyCy+99NJxj/Hyyy+zceNGlixZQlhYGLt27SIuLo5nn32W2bNnHzOinZeXx9///ne+/vprrLUMHTqUCy64gNjYWNauXctbb73F3/72N8aPH8/777/PD3/4w2OOmZ+fz1dffcX69esZOXIk69ato127dnz++edERkaydu1abrzxRg59E7lw4UK+/fbbmhHsTz/9tKa2vXv34vV6CQsLY9asWTz22GO8//77NcdKTU3lrrvuOmJEe8KECfzsZz9jxIgRbNq0idGjR7Nq1apT/j6k6akKVLFsxzKyfdnsr9zPL4f9ErfLTYQ7grFdx5KVlEVGQgaRYZFOlypNkIK2iIiDOnbsSGZmcPGKH/7wh7zwwgs1YfH6668HYN++fcybN4/rrvv+vPCDBw8CMHfu3JrQefPNN/PII8cuojtr1izuuusuwsKCb/lxcXEnrSknJ4err76a6OhoAK655hqys7O58sor6dy5MwMGDABg0KBB5OfnH/cxxo8fj8vlolu3bnTp0oXVq1fTuXNn7r33XpYuXYrb7WbNmjU1+w8ZMuSE00T27NnDrbfeytq1azHGUFlZedL6Dz3nlStX1mzv3buX0tJSWrZsecr7StMwr3AeH6z7gLlb5rK3Yi9hJoyhHYZircUYw39f8N9OlyjnAAVtEZFqJxuBbh7W/KS3x0bGhjSCfbSjT6Y6fPtQ0A0EArRu3ZqlS5eG9BhHOxQsQnWyc3eaNWtW87Pb7T7u1JHj1WSM4bnnnqN9+/YsW7aMQCBAZOT3I4iHnuvxPP7444wcOZIZM2aQn5/PhRdeeMrnEAgEmD9/Ps2bNz/lvtL4WWtZvWs12YXZ3NTrJqLDo/m2+Fu+3vY1IzuOxJPsYXiH4bSM0ActqV+aoy0i4qBNmzYxf/58AN566y1GjBhxzD4xMTF07tyZ6dOnA8FQsWzZMgAyMzN5++23AZg2bdpxj3HppZcyZcoUqqqCHRV27Qo2cmrZsiWlpaXH7O/xePjggw8oKytj//79zJgxg6ysrNN6XtOnTycQCLB+/Xo2bNhAjx492LNnD4mJibhcLt544w38fn9Ij7Vnzx6SkpIAeP3114+7z9HP5dJLL+XPf/5zzfaJPqRI43Wg6gBfbPqCJ+Y9wcXTL2b8x+N5ccmLLN+5HIBbet/C7PGzeWrEU1yaeqlCtjhCQVtExEG9evVi6tSppKWlsWvXLu6+++7j7jdt2jReffVV+vfvT58+fZg5M7iO15/+9CdeeuklMjIy2LNnz3Hve8cdd5CSkkJaWhr9+/fnzTffBGDSpElcfvnlNSdDHpKens7EiRMZMmQIQ4cO5Y477mDgwIGn9bx69OjBBRdcwOWXX86UKVOIjIzknnvuYerUqQwbNow1a9acdBT7cD//+c/5xS9+QWZm5gnD+RVXXMGMGTNqToZ84YUXyM3NJS0tjd69ezNlypTTql8apoK9BRTsLQBg095NPDD7Af6d/2/6t+vPbzJ/w+zxsxmWGFw4JjIsEpdRzBFnqb2fiJyznG7tlp+fz5gxY/j2228dq6EuTJw4kTFjxpyyj3dD4PRrQE6u0l9JblEuXp+X7MJsCvYWMLbrWH6T+RusteQV5dG/XX/CXeFOlyrnkNNp76c52iIiItJg7K/cT3R48NuO8R+PZ93udUS4IshIzGBCzwl4kj1AcN7/4ISQso6IYxS0RUQckpqa2uRGs+HE86hFjscf8LN853KyC7PJ9mVTVFbE7PGzcRkXk9ImERUWRUZCBlHhUU6XKnLaFLRFRETEER+t/4g/LPoDJQdLcBkXA9oO4ObeN1PhryAyLJLLO1/udIkiZ0VBW0TOaafb+k6ajqZyjlJjYK1l7e61wbnWvmweGvwQaW3TSIhOIDMpE0+yh/M7nE+rZq2cLlWkViloi8g5KzIykuLiYuLj4xW2zzHWWoqLi4/o5S21b3f5bl5c8iLZhdls3b8VgJ5xPdlfuR+AjIQMMhIynCxRpE4paIvIOSs5ORmfz8eOHTucLkUcEBkZSXJystNlNCm+Uh9en5cWES248rwriQqP4svNX5LWJo0fp/2YEUkjaB/d3ukyReqNgraInLPCw8NPuOy3iIRmyfYlfLnpS7w+Lxv2bADgwo4XcuV5VxLhjmDWuFm4XW6HqxRxhoK2iIiIhGzngZ0s2b6ESzpdAsDUFVP5yvcVg9sPZlz3cXiSPXSK6VSzv0K2nMscX7DGGJMPlAJ+oMpaO9gYEwe8A6QC+cB4a23JyR5HC9aIiIjUvoANsKp4FV6fF6/Py7fFwZaUn137GYktEtm6bysxzWJqel+LNHWNccGakdbanYdtPwp8Ya19xhjzaPX2I86UJiIicm4prSjFZVxEh0fzvxv+l8dyHsNgSGubxr0D7sWT7CEhOgGAxBaJDlcr0nA1lKB9tKuAC6t/ngrMQUFbRESkTlhr2bhnY3DUutDLkqIl/HzIz7mx541kJmXy2xG/ZUTSCGIjY50uVaRRaQhB2wKfGWMs8Fdr7ctAe2vtVgBr7VZjTDtHKxQREWliDvWQP+g/yDUzr2FT6SYAusV245Y+t5DeLh2AuMg4rjjvCidLFWm0GkLQzrTWbqkO058bY1aHekdjzCRgEkBKSkpd1SciItIkbNu/rWaudZgrjOdHPk8zdzNGdhxJSkwKWUlZmgoiUoscD9rW2i3Vl9uNMTOAIUCRMSaxejQ7Edh+gvu+DLwMwZMh66tmERGRxuS9Ne/x5uo3WVuyFoCkFklcnHJxze0PZzzsVGkiTZqjQdsYEw24rLWl1T9fCjwJfAjcCjxTfTnTuSpFREQaj5LyEnIKc8gpzGHy8MlEh0ezt2Ivsc1ieWjQQ3iSPXRu1VmroYrUA6dHtNsDM6r/2MOAN621nxpjFgHvGmNuBzYB1zlYo4iISIO2vWw7H6z7AK/Pyzc7vsFiiYuMI39vPn3i+3Bb39u4re9tTpcpcs5xNGhbazcA/Y9zfTFwUf1XJCIi0vCVVZYxf+t8EqIS6NOmDyXlJby45EX6xvflrv534Un20Du+Ny7jcrpUkXOa0yPaIiIiEoKCvQV4fV6yfdnkFuVSGahkXPdx9GnTh+6x3Zk9fjZtmrdxukwROYyCtoiISANU4a9gc+lmzmt9HtZaJn02iS37t9ClVRdu6nUTWUlZDGw/EABjjEK2SAOkoC0iItJAbC/bTrYvG6/Py4KtC4gMi2T2+Nm4jIunRjxFYnQiyS2TnS5TREKkoC0iIuIQf8CPMQaXcfHK8lf40+I/AZAQncCYLmPISs7CWgsGMhIyHK5WRE6XgraIiEg92nNwD/O2zMPr8zK3cC4vXvQi/dv2JyMhg/vT78eT7KFb625qvyfSBChoi4iI1IMt+7bwi+xfsGzHMvzWT+tmrRmRNIJIdyQA/dv2p3/bYxpxiUgjpqAtIiJSyw5UHWDh1oV4fV46xXTilj630KZ5GwI2wO39bicrKYt+bfrhdrmdLlVE6pCCtoiISC2ZsXYG/y74N4u2LqIiUEHzsOaM7z4egAh3BG/84A2HKxSR+qSgLSIicgYqA5Us3b6U5TuX16y6+JXvK3ylPsb3GI8n2cOg9oOIcEc4XKmIOEVBW0REJES7ynfh9Xnx+rzM3zKffZX7CHeFM7brWOIi4/ivrP+ieVhzp8sUkQYipKBtjGkP/BboYK293BjTGxhurX21TqsTERFxUMAGWFm8kg4tOhAXGYfX5+XxuY/Trnk7RqeOJis5i2GJw4gOjwZQyBaRI4Q6ov068Hfgl9Xba4B3AAVtERFpUkorSpm/ZT5en5ecwhyKy4t5bOhj3NjzRkaljKJHbA96xvVU+z0ROaVQg3Yba+27xphfAFhrq4wx/jqsS0REpF5YaymrKiM6PJp9Ffu48J0LqQhU0DKiJSM6jCArOYsRSSMAiImIISY+xuGKRaSxCDVo7zfGxAMWwBgzDNhTZ1WJiIjUoYP+gyzatqhmvnWnmE789ZK/0iKiBQ8NfoiecT1Ja5tGmEunMonImQv1HeRB4EPgPGPMXKAtcF2dVSUiIlJHns97nmmrplHuLyfSHcnQxKFclHJRze0Tek1wsDoRaUpCDdorgAuAHoABvgNcdVWUiIjI2aoKVLFsx7KaDiGvX/Y6UeFRJLVM4upuV+NJ9jC4/WAiwyKdLlVEmqhQg/Z8a206wcANgDFmMZBeJ1WJiIicobUla/nbN39j7pa57K3YS5gJY2D7gRQfKCYqPIrruusLWRGpHycN2saYBCAJaG6MGUhwNBsgBoiqrSKMMW4gFyi01o4xxnQG3gbigMXAzdbaito6noiINA3WWlbvWo3X5yW9fToZCRkEbICF2xYysuNIPMkehncYTsuIlk6XKiLnoFONaI8GJgLJwLOHXV8KPFaLddwPrCIY4AF+BzxnrX3bGDMFuB34n1o8noiINFIBG2D25tlk+7LJ9mWz/cB2AO7pfw8ZCRl0j+3Ol+O/xGU0w1FEnHXSoG2tnQpMNcZca619vy4KMMYkA/8BPA08aIKNSUcBh85GmQo8gYK2iMg5q2BvAYWlhZyfdD4Gw+8W/o7SilKGdxiOJ9nDiKQRtGneBgBjDAb1uBYR551q6sgPrbX/BFKNMQ8efbu19tnj3O10PQ/8HDj0vV48sNtaW1W97SM4feV49U0CJgGkpKTUQikiItIQVPoryS3Kxevzkl2YTcHeAuIi45g9fjYu4+Jvl/6NDi06EO4Kd7pUEZETOtXUkejqyxZ1cXBjzBhgu7U2zxhz4aGrj7OrPd79rbUvAy8DDB48+Lj7iIhI47C9bDvxkfG4XW6ezXuWf676JxGuCDISM5jQcwJZyVk100E6xXRyuFoRkVM71dSRv1Zf/rqOjp8JXGmM+QEQSXCO9vNAa2NMWPWodjKwpY6OLyIiDvEH/CzfubxmqfNVu1bxxuVvMKDdAK7pdg3DEocxJHEIzcOaO12qiMgZOdXUkRdOdru19r6zObi19hfAL6qPdSHwsLX2JmPMdGAcwc4jtwIzz+Y4IiLSMFhrMcawtmQtt//7dkoOluAyLga0HcD96feTGJ0IQLfYbnSL7eZwtSIiZ+dUU0fyqi8zgd4v7yglAAAgAElEQVTAO9Xb1x12W114BHjbGPMUsAR4tQ6PJSIidcRay9rda4NzrX3ZZCRkcO/Ae+kU0wlPsofMpEzO73A+rZq1crpUEZFaF0rXEYwxE4GR1trK6u0pwGe1WYi1dg4wp/rnDcCQ2nx8ERGpX8/mPsun+Z+ydf9WAHrG9aRdVDsAItwRPDXiKSfLExGpc6GuDNmBYFeQXdXbLaqvExERwVfqw+vzsn73eh4f/jgAxeXF9IrrxY/TfsyIpBG0j27vcJUiIvUr1KD9DLDEGDO7evsCgr2tRUTkHLWmZA0frf8Ir8/Lhj0bgGA3kP2V+4kOj+bpEU87XKGIiLNCCtrW2r8bYz4BhlZf9ai1dlvdlSUiIg3NzgM7ySnMYXjicNpHt2fFzhVMWzWNwe0HM677ODzJHrXdExE5TEhBu3q1xouBLtbaJ40xKcaYIdbahXVbnoiIOCVgA6wsXonX58Xr87KieAUAvxr+K8Z1H8dlnS9jdOpoosKjHK5URKRhOmHQNsZkAgustX7gL0CA4NLoTwKlwPtARn0UKSIi9aO0opTd5bvpGNOR3Qd3M+F/JwCQ1jaNnw78KZ5kDz1iewCov7WIyCmcbETbAv9DcInzodbadGPMEgBrbYkxJqI+ChQRkbpjrWXjno3BUetCL0uKljAkcQh/veSvxEXG8eeL/ky/Nv2IjYx1ulQRkUbnhEHbWjvPGFNWvVlpjHFTvRS6MaYtwRFuERFpZKoCVYS5gm//D331EJ8XfA4EF4m5tc+tXNjxwpp9PckeJ0oUEWkSTtVHe2n1jy8AM4B2xpinCa7a+P/quDYREakl2/Zvq5lrvXj7Yj4f9znR4dGMTh3NsMRhZCVlkdgi0ekyRUSalFC7jkwzxuQBFwEGGGutXVWnlYmIyFlbtG0R/7Xwv1hbshaApBZJjOkyhgNVB2qCtoiI1I1TBm1jjAv4xlrbF1hd9yWJiMiZKCkvIacwh2xfNmPOG4Mn2UPrZq2JbRbLw4MfJis5i84xnQk2khIRkbp2yqBtrQ0YY5YZY1KstZvqoygREQlNpb+S1759DW+hl+U7lmOxxEfGM7zDcCA47/rV0a86XKWIyLkp1JUhE4EVxpiFwP5DV1prr6yTqkRE5Lj2V+5nwdYF7D24l6u7XU2YK4wZ62YQ2yyWu/vfjSfZQ6/4XriMy+lSRUTOeaEG7V/XaRUiInJCm0s3M2fzHLw+L3lFeVQGKklpmcLYrmMxxvDBVR8QGRbpdJkiInKUUE+G/MoYkwAMIdjib5GWYBcRqRsV/gryivIYkjAEt8vNGyvf4K3Vb9GlVRdu6nUTnmQPA9oNqJlrrZAtItIwGWvtqXcy5g5gMvAlwa4jFwBPWmtfq9vyQjd48GCbm5vrdBkiImdke9l2sn3ZeH1eFmxdQFlVGf/8wT/p37Y/hfsKsdaS3DLZ6TJFRM55xpg8a+3gUPYNderIfwIDrbXF1QeIB+YBDSZoi4g0Jv6An4P+g0SFR7Fk+xJu+eQWABKiExjTJdgxpHtsdyDYkk9ERBqfUIO2Dyg9bLsU2Hy2BzfGRAJeoFl1Le9Za39ljOkMvA3EAYuBm621FWd7PBERJ+05uId5W+bh9XmZWziXa7tfy/3p99Mnvg8/G/QzspKy6Nq6q9rviYg0EaEG7ULga2PMTIJztK8CFhpjHgSw1j57hsc/CIyy1u4zxoQDOcaYT4AHgeestW8bY6YAtwP/c4bHEBFx3D2z7mHulrkEbIDYZrGMSBpBRvsMACLcEdzW9zaHKxQRkdoWatBeX/3fITOrL1uezcFtcIL4vurN8Or/LDAKmFB9/VTgCRS0RaQROFB1gIVbF+L1edmyfwv/c3HwratbbDd6xffCk+yhb3xf3C63w5WKiEhdC7XrSJ219zPGuIE8oCvwEsFAv9taW1W9iw/QBEURadDmFs7ljVVvsGjrIioCFUSFRTG8w3Aq/BVEuCP42aCfOV2iiIjUs1BHtOuMtdYPDDDGtAZmAL2Ot9vx7muMmQRMAkhJSamzGkVEDlcZqGRJ0RK8Pi8Tek2gQ4sO7DiwA1+pj/E9xuNJ9jCo/SAi3BFOlyoiIg5yPGgfYq3dbYyZAwwDWhtjwqpHtZOBLSe4z8vAyxBs71dftYrIuaessozPCj7D6/Myf8t89lXuI9wVzsB2A+nQogNXnnclY7uOdbpMERFpQE4ZtKundtxnrX2utg9ujGkLVFaH7ObAxcDvgNnAOIKdR27l+znhIiL1ImADrCxeScAGSGubRrm/nMlzJ9O2eVtGp44mKzmL4YnDiQqPAtCS5yIicoxTBm1rrd8YcxVQ60EbSASmVod5F/CutfZjY8xK4G1jzFPAEuDVOji2iMgRSitKa9rv5RTmsKt8F5kdMplyyRTiIuOYOXYmqTGpar8nIiIhCXXqyFxjzJ+Bd4D9h6601i4+m4Nba78BBh7n+g0El3sXEakz1lqKyopIiE4A4N4v7mXx9sXERMSQmZSJJ9lDZofMmv07t+rsVKkiItIIhRq0z6++fPKw6w614RMRaTTKq8pZtG0RXp+X7MJsisqKyLkhh+jwaH4y4CeEucJIa5tGmKvBnMIiIiKNVKjt/UbWdSEiInXFWosxhk/zP+XxnMcp95fTPKw5QxOGclvf2zAEp4IMSdQXaSIiUntCCtrGmMnHu95a++TxrhcRcVJVoIplO5bVjFrf3f9uLul0Cd1bd+fqblfjSfaQkZBBM3czp0sVEZEmLNTvRvcf9nMkMAZYVfvliIicubLKMn4171fM3TKX0opSwkwY6e3TiQoLdgbp0roLjw19zOEqRUTkXBHq1JE/Hr5tjPlv4MM6qUhEJATWWlbvWo3X58VlXNyZdifNw5qzZd8WRnUchSfZw/AOw2kZ0dLpUkVE5Bx1pmf7RAFdarMQEZFQzN8yn0/zPyXbl82OAzsA8CR7ADDGMO0/pjlZnoiISI1Q52gv5/tl0N1AW47sQCKHbF4I+dmQmgUdhxy7LSKnpWBvAXML53J9j+txu9zM3jybz/I/4/wO5wfb7yVl0qZ5G6fLFBFpeE6WQZRX6kWoI9pjDvu5CiiqXh5dDrd5IUy9EvwV4I6Ay56BTx/9fvvWD/XiFTmFCn8FuUW5ZPuy8fq8bCrdBMCAdgPoHd+bnwz4Cf+Z8Z+Eu8IdrlREpAE7OpMcnkGUV+pNSGsGW2sLgI7AKGttIdDaGKOVG46Wnx18kVp/8HLVzCO387OdrlCkQSraX8SOsuA0kNyiXH78+Y9597t3SYlJ4bGhj/HJNZ/QO743AK2atVLIFhE5laMzyeEZRHml3oQ6deRXwGCgB/B3IAL4J5B5svudc1Kzgp8ED30i7HUVFMz/fjs1y+kKRRoEf8DP8p3La9rvrd61mjv63cH96fczuP1g/jzqzwxJHELzsOZOlyoi0jgdnUkOzyDKK/XGWGtPvZMxSwkulb7YWjuw+rpvrLVpdVxfyAYPHmxzc3OdLkNznkROoDJQSbgrnIAN8IN//YDCfYW4jZv+bfvjSfYwKmWUljgXEalNmqNdJ4wxedbawSHtG2LQXmitHWKMWWytTTfGRAPzFbRF5ESstawpWUN2YTbZvmzKqsqYfsV0AKatmkZcZBzndzifVs1aOVypiIhI6E4naId6MuS7xpi/EpybfSdwG/C3My1QRJq299e8z5RvprBt/zYAesX1wpPswR/w43a5uanXTQ5XKCIiUvdCXbDmv40xlwB7Cc7Tnmyt/bxOKxORRmFz6eZgh5BCL48Pe5ykFklEh0fTJ74Pd6XdRVZyFu2i2jldpoiISL0LecGa6mCtcC0i7CjbwdQVU/EWetm4ZyMAqTGpFO0vIqlFEpd1vozLOl/mcJUiIiLOOmnQNsaU8v1CNUfcBFhrbUydVCUiDcrOAzvJKcwhPjKerOQs3C43b3/3Nunt0hnffTyeZA8pMSlOlykiItKgnDRoW2tb1lchItKwrNi5gq98X+H1eVlRvAKAy1MvJys5i7jIOHJuyCEyLNLhKkVERBqukKeOABhj2gE1/7JaazedzcGNMR2BfwAJQAB42Vr7J2NMHPAOkArkA+OttSVncywRObnSilJW71pNRkIGAP+18L/4Zsc3pLVN46cDf4on2UOP2B41+ytki4iInFyo7f2uBP4IdAC2A52AVdbaPmd1cGMSgURr7WJjTEsgDxgLTAR2WWufMcY8CsRaax852WOpvZ/I6bHWsmHPhpoTGZcULQEDOTfkEB0ezbqSdcQ3jyc2MtbpUkVERBqMumjv9xtgGDDLWjvQGDMSuPFMCzzEWrsV2Fr9c6kxZhWQBFwFXFi921RgDnDSoC0ip1ZeVY7LuIhwR/Dm6jd5ZuEzAHSP7c6tfW7Fk+wh0h0cqe4a29XJUkVERBq9UIN2pbW22BjjMsa4rLWzjTG/q81CjDGpBFef/BpoXx3CsdZurZ6ycrz7TAImAaSk6EQskePZum9rzVLnX2/9mt9m/ZZLOl1CVlIW4cPC8SR7SIhOcLpMERGRJifUoL3bGNMC8ALTjDHbgaraKqL6sd8HHrDW7jXGhHQ/a+3LwMsQnDpSW/WINAW7y3fzo3//iHW71wGQ1CKJq7tdTUrL4IfSlJgUdQoRERGpQ6EG7auAA8DPgJuAVsCTtVGAMSacYMieZq39V/XVRcaYxOrR7ESC88JF5AR2le9ibuFcsn3ZxDeP55Ehj9CqWSu6te7G2K5jyUrOonNMZ0L9ECsiIo1DXkEJCzYUExsVQUlZRc3lsC7xDOqkc2ycFmrQngRMt9b6CM6ZrhUm+K/+qwRPrHz2sJs+BG4Fnqm+nFlbxxRpSt797l1mrp/J8h3LsVjiI+O54rwrADDG8PsLfu9whSIiUlfyCkq46ZUFHKwMYKle5KT6slm4i2l3DGNQp9iaMK7wXf9CDdoxwL+NMbuAt4H3rLVFtXD8TOBmYLkxZmn1dY8RDNjvGmNuBzYB19XCsUQatf2V+1mwZQFfb/uaRzIewe1ys373egKBAHf3vxtPsode8b1wGZfTpYqISD14f7GvJmQDR1xWVgVYsKEYgJteWUBFVYCIsO/Dt9SPkIK2tfbXwK+NMWnA9cBXxhiftfbiszm4tTaH4Aev47nobB7bSfrkKLVle9l2Pt34KdmF2eQW5VIVqKJFeAtu6nUTnWI68ciQRxSsRUTOQXkFJbyX5zvu8t0AbpdhWJd4FmwopqIqQMAGw/f7i33KKPXotBasIThXehtQDBy3E8i57tDXOPrkKGeiwl9BblEuyS2SSYlJYV3JOv6Q+we6tOrCD3v9EE+yhwHtBhDuCgdQyBYROUct2FBMlT8ABEcsz2sbzfod+2umjlw3uGNN/ogIc1FZFcDtMrybuxm/3xLuNrw1abgySh0LKWgbY+4mOJLdFngPuNNau7IuC2usjv7kuGBDsV7ETVhtfHtRtL+I7MJssn3ZzN86nwNVB7iz353cl34fGQkZfHLNJyS3TK7lykVEpDEb1iW+JkCHh7m4bUQXnvx4Rc32NenBfzcGdYpl8pg+vLNoEzv3V1BYcgCACr/l/cU+ZZQ6FuqIdieCrfeWnnLPc9zRL/xhXeI1laSJyiso4caX51Ppt7hdcH1GCtekJ5/wd5xXUMK/FvsI2ACj+kRySY8eVAWqGDtzLPsq95EYnciV512JJ9lTswx6uDtcIVtERI4xqFMs0+4YdkS+6JHQ8pi8kVdQwhMfraCiKnDMY6gPVd0LaQn2xqChLMF+KExZ4Nr0ZL7bVsrkmd/iD9gjzgCWxm/SP3L5bOWR5wRHnuB3/Oq8Ffzuq5m4olfhjl4D/hj+MfpdBnWKZVbBLDrFdKJr665qvyciIrUmr6CE52etIXvtzmNuiwhz8dadyiRnoi6WYJcQHD0/u2+HVkye+S1VgeCHmYpKTSVpKvIKSvhi1bGNdw5NF0pPaV0Tmh+b8wc+zH+DZh0sgapo/Pt6ULWvJ/PX72RQp1gu7nRW5xSLiIgc4/BMcriwEL6BldqjoF2LDp+ffbAywAtfrMEf+P4bA1f1GcDS+C3YUIz/qC+DjKnA3XI9U7/7iNc2reL2Ln/AVsWxfUd7KotHUlnak0B5MhA8gTEuuln9Fy4iIueEwzOJy0C/pFb0SWrFtdUBO6+ghJdmr6uZZqJprnVDQbsWXdQin9iw1+hifcSZUjaUJbLBnUAfVwGfBobQKvNOgCNe2NL4rF40i67ffcYEdxUTzCwimm3jj21iyI2MpMJlMIEA5++rYOjsa+ldtZ8IAoDFAgfDw4kwfvbY5mR/8xMY+kvYvBDysyE1CzoOcfrpiYhIE3D0OWOTr+hzxLztw7+BnzymD09+vEId0+qAgnZt2byQnp9OoIf7YE3H+G4U1tzscS3nl/MM1+eMImDRC7mRWr7wU3Z/cTvfRIWR2aqCPvvL2BVwsSUsjPGlpXjKDpB+4CARh+5w1JTrKFMJQLzZz1jf79n6/l4SV00FfwW4I+DWDxW2RUTkrB3vZMlD/nXYQjeVVQE++XarOqbVEQXt2pKfDf6KYK46LFxZC8YEL0ebr3kzMAo4csUmfVXT8H24/kPmbJ5DdsFsyjvEEW4tt+zeiymDeBvgo8Kt3+8c4vmM1kLk2v8NhmzrD17mZytoi4hIrRjUKfaYbJFXUML03M01C9243S4u75vIovxdR3RMk9qhoF1bUrPAHYH1fz+ifcihxi6f+A8LUMawtqiU5z5fQ8BawlyGC3u0o03LZjXzp8QZARtgZfFK1u9ez1VdrwLg3e/eZeu+rWS2Hszo1Z+QeWA/LW0AeyhUn2HznvJu/wGHj2inZtXOkxARETmOBRuKa5o0GGDcoGQmDE05bmtAOXtq71ebNi9ke87rbFyZRyylbLSJrLfBOdqf+IfwdiC0VeUjtFpTvSutKGXelnl4fV5yCnPYVb6LZu5m5NyQQ2RYJCXlJbRuFuwksnrRLEpWfkliYhKpG96C7avwB6owh53YfZBwymwEUaaccAKYQ3O0bTgRbj8HTDS7hv+C1Et/ojnaIiJSbw7Nzz40eq1prKfvdNr7KWjXspdmr+OPn31H4Cz/t04YmsJvr+5XO0XJMay1bNyzkYToBKLCo3hl+Sv8afGfiImIITMpE0+yh8wOmcRGnvrNJ6+ghOv+Zx7HLgVwfAa09K2IiDhGHUbOjvpoOySvoIRlm3djCIap08naLsMR4Xx67mZNIall5VXlLNq2CK/PS3ZhNoX7Cnn2wme5pNMljOkyhvR26aS1TSPMFfqfxaHFAEIN2RB8XVT4LU9+tOKIs8BFRETqw/Hmbkvd0Ih2LckrKOHGvx3bGD5UE4amsHDjLtZt31dz3aW92/PyLSF9YJITqAxUEu4KZ+u+rVz5wZWU+8tpHtacoQlDyUrOYlTKKNo0b3NGj33o67dDZ26fiROtJCkiIiINk0a0HXCoMfyZcBno26EVwBFB+4tVReQVlCiEnYaqQBXLdizD6/Pi9XnpHd+bp0c8TUJ0Ajf3vpn09ulkJGTQzH32i8Uc+p2fzUdVtVESERFpuhS0a8mwLvHHTP8IVcDC5JnfcseIzrgNNSsOWkvIIUzzreC/F/03/1r3L0orSgkzYaS3Tye9XToAxhjuS7+vVo93aDGAE41ohzJ9SG2UREREmi5Hg7Yx5jVgDLDdWtu3+ro44B0gFcgHxltrS5yqMVSDOsUyuFMsC/PPrNSqgOWVnI3cmdWFV3I2EghYIsJDC2FHr/DU1KciWGtZtWsV2b5sluxYwkujXsLtchMVHsWojqPwJHsY3mE4LSNa1mkdgzrFMnlMH/7fB8s53gysUD5zTRyeetLflT5AiYiINF5Oj2i/DvwZ+Mdh1z0KfGGtfcYY82j19iMO1HbaurZvecZBG4Jhe/GmEt758fDTCleHpjA09RWdvtv1HW+ufpNsXzY7DuzAYOjbpi/F5cW0i2rHPQPuqfeaSsoqjhuyQzV/QzGPzViOAa456uTXc+0DlIiISFPjaNC21nqNMalHXX0VcGH1z1OBOTSSoH1tejLv5W6mwn/myWthfgmfr9jGoz/oFfJ9Dk1haGorOuXvySe7MJuhiUPpHtud4vJiPsv/jPM7nI8n2cOIpBHEN3f2ucZGRZx05NoFJ+1I8u2WPSzz7QFgep6Pt+78PkyfKx+gREREmiqnR7SPp721diuAtXarMaad0wWFalCnWJ64si/ffvQnRpuvWRHoRIw5AMC//Fkstt1Depy3czdzSZ+EkEPVoE6xTLtjWKOfYlAVqGLhtoVk+7Lx+rxsKt0EwEODHqJ7bHeGJAzBe4OXcFe4w5V+b85324+5Lt2s4Rp3Nm3YQ7uYZqzcG3nE7z/drGG4exW72w3lzS0JNfc7OkwP6xLPkLB1DLIryDN9GNbl/Pp5UiIiIlIrGmLQDpkxZhIwCSAlJcXhaoIS1r3Nje5XAPC4ltdcf537K26s+H8hhe3dZZXc+PL801rQpLH2xCzaX8SOAzvo26Yv1loenPMgVYEqMhIy+GHvH5KVlEVyy2SA0+pvXV+K9pYfsZ1u1vBWxFNEUBW8ogwGur///QNMi/gt4VQRKPmAtWGPsbCqG3DsiZGDXGt5M+K31cuzz8TlGg5o5UgREZHGouElFygyxiRWj2YnAscOGVaz1r4MvAzBPtr1VeDJpBbNAsCYYNcQY4LXh1k/w1yrWOwPbVS7wm95f7GvUYbnk/EH/Czfubxm0ZjVu1bTtXVXZlw1g3B3OK9e+ipdWneheVhzp0sNyfUZKSzzff+BaphrFWH4a37vAJjvf/8A4VQRZgJYqnhu6D5eqko57hxt8rNxBSqBAAQqg8u0a4l2ERGRRqMhBu0PgVuBZ6ovZzpbTujyCkp4v7gvT4d9XXOC3KHLKtwsCIQ+7xqC7eGOd4zGNkWktKK0pgPI5HmT+XD9h7iNm/5t+/NA+gN4kj01+/Zp08epMs/IhKHBb1I++XYrfRJj6FJxGYGl/8JlK4/Y79DvPy46goD/AyxVGHcESQMu5bcd+x3/wVOzwB1RPaIdEdwWERGRRsPRlSGNMW8RPPGxDVAE/Ar4AHgXSAE2AddZa3ed6rGcXhkS4LEZy3nz603c4PqCy90Lz3iO9iF3ebrQsnl4TahuLF0orLWsKVlDdmE22b5slu1YxsdXf0xyy2TyivLYUbaD4R2G06pZK6dLrROrF81i8Yd/Ic4GT3LcSasjfv/DwtfxxyGlJA249NQj1JsXBkeyU7M0mi0iItIANJqVIa21N57gpovqtZBacmgE+u3ARbwdOP2nkBDTjKK9B7HVj/VKzkYC1taE6sbQhWLZjmU8/NXDbNu/DYBecb24vd/tRLgjABjUfpCT5dWLnhkXs7/dIN5f7GNn6UF2l1WwZVcZ7D0IwKKqrnzQogc/6dj11A/WcYgCtoiISCPVEKeONFrXpCezJu8LrsLLcLOCZLMDi6EKF5FUYoH1gUSKTDyf+IfUhPF0s4ZhrlVEJHr4y/44Kv22epVJe0Sobmht/DaXbq6Zaz2q4yjG9xhPx5Yd6RPfh7v7382IpBG0i2o0TWNq1aBOsQzaMRNWzYQoOLh3JesimmEtxJtS2DUe+J1GrEVERJowBe1aNMi1lncinsYEKoJXVM/KaXbYPj3cW+jBlmBHkkpYYzvWdKGozJ9BDr8kl241o9ou8303iobQxs9ay3OLn2PO5jls3LMRgNSYVFzGBUBcZBzPj3y+3utqcHJfh4/vr9lsBvR2fX+zWT4F2EVg5Yc1c7BdEz9S2BYREWlCFLRrU02XiGrHOZvxUCcSa+Fy90LiAvtqulBgqxjCSnLpRqA6pIe5DJPH9KkJ1fXdxm/ngZ1k+7IpKivirv53YYzh253fkhCVwPju4/Eke0iJaRitFRuUVceew3v0y6Hyu88wVRWEmQBVVf+/vXsPj6q+8zj+/s4kIwnXlKgo4foICNYriLFcVpSy2rWi1WqVfdZefNxaanW3dtvafbrW1t6vf7hdfWi37vN4qau1svSCl2qhWEACWMSIRRSJICgGBAJJZua7f5yTMCGTEJJzmBn6eT0PT87vN+ec+c1vwplvfvM9v18L29Y+wXAF2iIiIscMBdpRGj0DkuV4poXulgtsu//0d5mpvOIjaKUMPE0rZZ1mJslmnd+9uI0JwwZ2GWBHPRPJxsaNLN68mKUNS1m/cz0AwwcM54bTb6AsUcaCOQvaR7ClCxPnwqt/6HaXjYPPZ/SOp9vf+z9nJnHVUWqeiIiIxE+BdpRGTOXHNT+i+tVfHVGO9ryW26lN1LM8O7HTzCRZYOlf32HFa+92WJ67TRQzkexp2cNzW59j2snTGJAawLMNz3LPC/dwxvFncPPZNzOzZiYTqiZg4eTQCrJ7YMrHg5/1jwd/dL3zMlSGOfVNO+H0q2kafyufWPA/7Ss/fuHsWYVqrYiIiMSgoNP7RakYpvf79m/r+a8lmyI7n9FxYHzeeSO564qOcy7f/cxGfvDEBrIOSYN/nTOB+bO6n83C3dm0e1P7jYxrtq8h7Wl+eMEP+eCoD7LrwC4cp6pfcc1ociwqxXnRRURE/paVzPR+x5K6zY3cuzR/kH2OvcJHkksBeM8ruCixBgOeyp7NXvrnHcmGztkn+f4k6ulMJAfSB9jbupfqimpee+81Ln/8cgDGV43n+tOuZ2bNTM44/gwAhvQb0qPXLH13tHPuRURE5OhRoB2R5Zt2ku/LgXPsFR5MfYMU6U6PjUu8SRajhXLmtdze7YI2SYMrz6npVN/dTCTb9m5jScMSlry5hJXbVjJn9Bzumn4XYwaN4RvTvsF5J53HsP7DeveCRURERKRbCrQjUjt2KOVJoyXTMdquTdRTTpowvbl91pG27aQ55Z6mNlHP6kzHQDthbT+NG6aPYfmmnQCdRkDbRkVz04DmPz2fJQ1LAKgZUMNHxn2E2aNmA2BmzD1lbiSvW0RERAjlw0YAABFdSURBVETyU6AdkcmjqvjolBHcv+KNDvXLsxNppYyUHxzRzh35zrjlnW0EYERVJVefO4KqyhR3Llqf94bHdw+8y7I3l7GkYQn179bz+NzHSSaSTB8+nanDpjKjZgZjBo1pv5FRRERERI4OBdoROu3kwZ3qVvt4rm35917laG9+t4k/btjBuBMHdlp6vbmsnrtfuJt1b6/DcYb2G8qMmhnsS+9jUGoQ157a1er2IiIiInI0KNCOUGNTS9761T6e1emDgfR3M9f1+JwrX29kbcN2UoM2YpUvwZ7zqR37AVoT75DNZrnpzJuYWTOTiUMnato9ERERkSKiQDtCVZWp6E5mLZRXraBswAaSla9hliFllXzi/EvCtJFaai+tje75RERERCRSCrQjtH7r7t4fbGmSla+BpcnsnQie4LjqJ8mmh5DdNZ2vzLqSq0+fTnmiPLoGi4iIiEhsFGhHpG5zI1PqvsjXUsvoaQLH9mSSP1X2Y2n/CpZX9GN/IsGkA8082LIdgMY3EgzOZsmWv0RqYzPUL4C31kFZP6j9zMHVB7eshBceBBzOvA5GTI3jJUqUtqyE15dCxVDYvxNGz9D7JiIiR0fbZ5A+e2KnQDsilYtu4vLksm73yQAbUuVMamkF4HvVVTw5oJKTWtN8eM8+Zjbt59wDze3T/73Ps2CQzDTBuoc7LFhji24JNk6cBL+4FDLNQXnN/fDx3+g/TjHbshLuuyx4zzwLJKDsOLh+od43ERGJV/tnUAskU/rsiZkC7YicsvvPwME5stvsTiRYVtGPJZUVLKvox65kkt9teZOadIZP79rNTbt2c0prKx0OyzMTnx9S7YDVPx6MhmZybsLMtAZ/peo/TfF6fWnwnnk2rMgGZb1vIiISt/bPoIw+e46Cog20zexi4CdAEljg7t8ucJO6VT5hDr7uYbJA2qEcWFrRj5uHHU/WjKpMhulNB5jZtJ+qdBZ3GBeObEP+5dXb5Sxw06Fu4txgRDuZOjiinSwPvgqS4jV6xsH3rG1EO5nS+yYiIvFr/wxq0WfPUWCeb93wAjOzJPAK8EGgAXgeuNbdX+rqmClTpviqVauOUgs7amptYuVbK1my5OssPbCVf9r1HvN27+XdRIIHBg9kZtN+TmtuIdmLc2eBdLKCrcMuZOOb2znVN9NsKcqnz2f0nPnBTsrRLj3K0RYRkUJRjnafmFmdu0/p0b5FGmifD9zh7n8flr8M4O7f6uqYQgTamWyGm/9wMyu2raAl20IqUcG+XWNp3XUemX2dF6A5UkawDLsDqbIEX730NBqbWqgdO7TTMuwiIiIiEr8jCbSLNXVkOLAlp9wAnHfoTmZ2I3AjwMiRI49Oy3IkE0kGHTeIqydczUnl53DnI/tJZ3ozbt3RkIpydu9vxYFM+HdQazpLY1ML82ed0ufzi4iIiEj8ijXQznM7YOc0Zne/F7gXghHtuBuVz7dnBKnjdz+zkWx2QyTn3NOc7vRiy8sS1I4dGsn5RURERCR+xbpmdwMwIqdcA2wtUFt6pHbsUFJl0XRnNtsxzE4mjK9eeprSRURERERKSLEG2s8D48xsjJmlgI8BCwvcpm5NHlXF/TfUMmNcdd7h+CORZ+iexqaWfLuKiIiISJEqykDb3dPAZ4HFQD3wsLuvL2yrDm/yqCpunT2esmRfQ+2DEqa0EREREZFSVKw52rj7b4HfFrodR6JucyO/Wt1AJhNNungCmHZKNbfOHs/kUVXUbW5k+aadmnVEREREpAQUbaBdauo2NzJvwXKaW7PdLz7TQwkLpvTLDbLnLVhOSzpLqizB/TfUKtgWERERKWIKtCOyfNNOWtJ9D7LPrBnMNeeO7DRfdtv5sx5M9bd8004F2iIiIiJFTIF2RNpmHWlNZ7GEkc062V5E3decO5Lrzus8J3ju+ZWzLSIiIlL8FGhHpG3WkeWbdrJ1134eWPFGj4475fj+bHx7HxCki3Q1u0ju+ZWjLSIiIlL8inLWkVI1eVQV82edwmknD+5RCknC4JPTx5JKGkawSk9VZeqw51eQLSIiIlL8FGjHoLGppUdzaV925slMGDYQh/bl1u9Y+CJ1mxu7PKZucyN3P7Ox231EREREpPCUOhKDqsoUyYSRyXq3I9u/XruVDW/toTVnOsDWjHd5o6NmHhEREREpHRrRjljd5kbuXLSerDsJC9JDulP/1p4O5fKkdXmjY76ZR0RERESkOCnQjlhuMAxw+vDBPT7WDO647P0AedND2mYeSWq1SBEREZGip9SRiFVVpkiYgTvlZQnGVPfnhYbdPTrWHZ7dsIM7F63Pmx6imUdERERESocC7Qi1pY1kso4ZXHzaMBb9ZdsRnePp+u04kHVobs3y6OqGDgH15FFVCrBFRERESoBSRyK0fNPO9iXYsw4LX9hK+ghXrck67TOWOPBIXYNmGBEREREpQQq0I1Q7dijJnLsf3YObIY3D3xTZJpEwLpp4Yns5ndFNjyIiIiKlSIF2hCaPquLOue+nLBEsQNM2sp0wSIZ1EATe/coTfHrmWM6sGUwy510wnLHV/dvLWe9+ERsRERERKU7K0Y7YdeeNZMKwgfz4qVf401/faQ+2s5lgTu0EMG1cNbfOHt+ea337Y+t4cMUbwcI1Duu3vdceqCfoell2ERERESleGtGOweRRVdw6ezzHlYdT8SWN8nBavlR5okOQDXDlOTUH9y1LcMn7T2ovp8o1jZ+IiIhIKTL3I7tZL7InNvsocAcwEZjq7qtyHvsy8CkgA3zO3Rcf7nxTpkzxVatWHW63o6puc2P7VHxAt9Py5e47eVRVp7KIiIiIFJ6Z1bn7lB7tW8BAeyKQBe4BbmsLtM1sEvAgMBU4GXgKGO/ume7OV4yBtoiIiIgcW44k0C5Y6oi717v7hjwPzQUecvdmd38N2EgQdIuIiIiIlIxizNEeDmzJKTeEdSIiIiIiJSPWWUfM7ClgWJ6HvuLuj3d1WJ66vPktZnYjcCPAyJEje9VGEREREZE4xBpou/vsXhzWAIzIKdcAW7s4/73AvRDkaPfiuUREREREYlGMqSMLgY+Z2XFmNgYYB6wscJtERERERI5IwQJtM7vCzBqA84HfmNliAHdfDzwMvAT8Hph/uBlHRERERESKTcGm94uamb0NbC7AU1cD7xTgef8WqG/jo76Nj/o2Purb+Khv46O+jU+h+naUux/fkx2PmUC7UMxsVU/nUpQjo76Nj/o2Purb+Khv46O+jY/6Nj6l0LfFmKMtIiIiIlLyFGiLiIiIiMRAgXbf3VvoBhzD1LfxUd/GR30bH/VtfNS38VHfxqfo+1Y52iIiIiIiMdCItoiIiIhIDBRo94GZXWxmG8xso5l9qdDtKWVm9nMz22FmL+bUvc/MnjSzv4Y/qwrZxlJkZiPM7Bkzqzez9WZ2S1ivvo2AmfUzs5Vm9kLYv18L68eY2Yqwf39pZqlCt7UUmVnSzNaY2aKwrH6NgJm9bmbrzGytma0K63RNiICZDTGzR8zs5fC6e776tu/MbEL4+9r27z0zu7UU+laBdi+ZWRK4G7gEmARca2aTCtuqkvYL4OJD6r4EPO3u44Cnw7IcmTTweXefCNQC88PfU/VtNJqBC939TOAs4GIzqwW+A/wo7N9G4FMFbGMpuwWozymrX6Mzy93PypkaTdeEaPwE+L27nwqcSfD7q77tI3ffEP6+ngVMBpqAxyiBvlWg3XtTgY3uvsndW4CHgLkFblPJcvclwLuHVM8F7gu37wMuP6qNOga4+zZ3Xx1u7yG46A9HfRsJD+wNi+XhPwcuBB4J69W/vWBmNcA/AAvCsqF+jZOuCX1kZoOAmcDPANy9xd13ob6N2kXAq+6+mRLoWwXavTcc2JJTbgjrJDonuvs2CAJG4IQCt6ekmdlo4GxgBerbyITpDWuBHcCTwKvALndPh7vo2tA7Pwb+DciG5aGoX6PiwBNmVmdmN4Z1uib03VjgbeC/w5SnBWbWH/Vt1D4GPBhuF33fKtDuPctTpylcpCiZ2QDgUeBWd3+v0O05lrh7Jvw6s4bgm66J+XY7uq0qbWZ2KbDD3etyq/Psqn7tnWnufg5B6uN8M5tZ6AYdI8qAc4CfuvvZwD6KMJWhlIX3ZVwG/G+h29JTCrR7rwEYkVOuAbYWqC3Hqu1mdhJA+HNHgdtTksysnCDIvt/dfxVWq28jFn5F/CxBLvwQMysLH9K14chNAy4zs9cJ0vIuJBjhVr9GwN23hj93EOS5TkXXhCg0AA3uviIsP0IQeKtvo3MJsNrdt4flou9bBdq99zwwLrwLPkXwVcbCArfpWLMQuD7cvh54vIBtKUlhXuvPgHp3/2HOQ+rbCJjZ8WY2JNyuAGYT5ME/A1wV7qb+PULu/mV3r3H30QTX1j+4+zzUr31mZv3NbGDbNjAHeBFdE/rM3d8CtpjZhLDqIuAl1LdRupaDaSNQAn2rBWv6wMw+RDDKkgR+7u53FbhJJcvMHgQuAKqB7cB/AL8GHgZGAm8AH3X3Q2+YlG6Y2XRgKbCOg7mutxPkaatv+8jMziC4ASdJMHDxsLvfaWZjCUZi3wesAf7R3ZsL19LSZWYXALe5+6Xq174L+/CxsFgGPODud5nZUHRN6DMzO4vgBt4UsAn4BOG1AfVtn5hZJcG9cWPdfXdYV/S/twq0RURERERioNQREREREZEYKNAWEREREYmBAm0RERERkRgo0BYRERERiYECbRERERGRGCjQFhEpMma29yg/31nhdKVt5QvM7AMRnv8OM7stqvOJiJQKBdoiIseonFUUD+cs4EM55QuAyAJtEZG/VQq0RURiZGbfMbPP5JTvMLPPm9kAM3vazFab2Tozm9vF8V8ws+fN7C9m9rWwbrSZvZizz21mdke4/ayZfdPM/gjccsi5pprZc2a2Jvw5IVzZ9k7gGjNba2ZfBD4N/EtYnhGugPlo2I7nzWxazmv5eficm8zscznP9RUz22BmTwETcuqfNbMp4XZ1uMw6ZpY0s+/lvNZ/7ku/i4gUg56OdoiISO88RLCC7H+G5auBi4EDwBXu/p6ZVQPLzWyh56wiZmZzgHHAVMCAhWY2k2AFtO4Mcfe/y1P/MjDT3dNmNhv4prtfaWZfBaa4+2fD560A9rr798PyA8CP3P1PZjYSWAxMDM95KjALGAhsMLOfAmcQLJ1+NsHnzGqg7jBt/hSw293PNbPjgGVm9oS7v3aY40REipYCbRGRGLn7GjM7wcxOBo4HGt39DTMrB74ZBs5ZYDhwIvBWzuFzwn9rwvIAgsD7cIH2L7uoHwzcZ2bjAAfKe/gyZgOTzKytPMjMBobbvwmXQW82sx3ha5gBPObuTQBmtrAHzzEHOMPMrspp6zhAgbaIlCwF2iIi8XsEuAoYRjDCDTCPIPCe7O6tYQpFv0OOM+Bb7n5Ph0qzGjqm/h163L4u2vF14Bl3v8LMRgPP9rD9CeB8d99/SDsAmnOqMhz8XHHyS3Ow7bntNuBmd1/cwzaJiBQ95WiLiMTvIYJUiqsIgm4IRmx3hEH2LGBUnuMWA580swEAZjbczE4AtgMnmNnQMM3i0h62YzDwZrj98Zz6PQSpH12VnwA+21Yws7MO8zxLgCvMrCIc+f5wzmOvA5PD7aty6hcDN4Uj/ZjZeDPrf5jnEREpagq0RURi5u7rCQLXN919W1h9PzDFzFYRjG6/nOe4J4AHgD+b2TqCIH2gu7cS3MC4AliU79gufBf4lpktA5I59c8QpIasNbNrgP8jCJTXmtkM4HNhW/9iZi8R3CzZ3etdTZC+shZ4FFia8/D3CQLq54DqnPoFwEvA6vBGz3vQt64iUuIs574bERERERGJiEa0RURERERioEBbRERERCQGCrRFRERERGKgQFtEREREJAYKtEVEREREYqBAW0REREQkBgq0RURERERioEBbRERERCQG/w8Tv2hKQtC/rAAAAABJRU5ErkJggg==\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["fig, ax = plt.subplots(1, 1, figsize=(12,4))\n", "ax.plot(y_test, clr.predict(X_test), '.', label=\"lin\u00e9aire\")\n", "ax.plot(y_test, clr2.predict(X_test), '.', label=\"arbre\")\n", "ax.plot([0,70], [0,70], '--', label=\"pr\u00e9diction parfaite\")\n", "ax.set_xlabel(\"valeur attendue\")\n", "ax.set_ylabel(\"valeur pr\u00e9dite\")\n", "ax.legend();"]}, {"cell_type": "markdown", "metadata": {}, "source": ["L'arbre de d\u00e9cision est nettement meilleur simplement parce que la relation entre la valeur \u00e0 pr\u00e9dire et les variables n'est pas lin\u00e9aire."]}, {"cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["from seaborn import pairplot\n", "xy = X.copy()\n", "xy['duree'] = y\n", "pairplot(xy);"]}, {"cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [{"data": {"text/html": ["\n", ""], "text/plain": [""]}, "execution_count": 21, "metadata": {}, "output_type": "execute_result"}], "source": ["from sklearn.tree import export_graphviz\n", "dot = export_graphviz(clr2, out_file=None, feature_names=X.columns)\n", "from jyquickhelper import RenderJsDot\n", "RenderJsDot(dot)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On voit bien que l'arbre de d\u00e9cision s\u00e9pare le d\u00e9but et la fin de la journ\u00e9e. Apr\u00e8s 17h30, la dur\u00e9e est augment\u00e9e de 8h en moyenne, si c'est vendredi, la dur\u00e9e est augment\u00e9e d'une vingtaine d'heure. Un mod\u00e8le lin\u00e9aire ne peut pas prendre cela en compte."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Remarque sur le score R2\n", "\n", "On voit que le score de pr\u00e9diction $R^2$ est nettement meilleur pour l'arbre de d\u00e9cision mais c'est un petit peu en trompe l'oeil. La dur\u00e9e de la p\u00e9riode d'inactivit\u00e9 joue un r\u00f4le non n\u00e9gligeable. Plus elle est longue, plus la dur\u00e9e entre la commande et la r\u00e9ception est grande en moyenne et c'est cette donn\u00e9e que le mod\u00e8le lin\u00e9aire n'arrive pas \u00e0 prendre en compte."]}, {"cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [{"name": "stderr", "output_type": "stream", "text": ["100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 24/24 [00:09<00:00, 2.83it/s]\n"]}], "source": ["from sklearn.model_selection import cross_val_score\n", "from tqdm import tqdm\n", "import numpy\n", "res = []\n", "for end in tqdm(range(1, 25)):\n", " scores_tree = []\n", " scores_lin = []\n", " for i in range(0, 5):\n", " df = duration_selling(hour_end=end, hour_begin=0, week_pattern=[1] * 7)\n", " data = df.copy()\n", " data[\"heure\"] = data.commande.dt.hour + data.commande.dt.minute / 60\n", " data[\"wk\"] = data.commande.dt.weekday\n", " data['duree'] = (data.reception - data.commande).dt.total_seconds() / 3600 \n", " X = data[[\"heure\", 'wk']]\n", " y = data[\"duree\"]\n", " clr = LinearRegression()\n", " clr2 = DecisionTreeRegressor(max_depth=3)\n", " score_tree = cross_val_score(clr2, X, y, cv=5).mean()\n", " score_lin = cross_val_score(clr, X, y, cv=5).mean()\n", " scores_tree.append(score_tree)\n", " scores_lin.append(score_lin)\n", " res.append(dict(end=end, score_tree=numpy.array(scores_tree).mean(),\n", " score_lin=numpy.array(scores_lin).mean()))\n", "res = pandas.DataFrame(res)"]}, {"cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [{"data": {"image/png": "iVBORw0KGgoAAAANSUhEUgAAAU0AAAElCAYAAABgV7DzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzsnXd4VFX6xz9veg8tCUnoECAJVbqwNhTBhu6Cirr2rrj7W8vqutbVta1tFV3rYqWIZcFGU0QQqdJbQodAQicJBFLO749zA0NITyaTmbyf57nPzL333HPee2fmO6e97xFjDIqiKErl8PO0AYqiKN6EiqaiKEoVUNFUFEWpAiqaiqIoVUBFU1EUpQqoaCqKolQBFc0GhogYEelQzWt/JyLratumMsraLCLn1kVZdYmIPCUie0RkVx2XmyMi7eqyTF9FRbOe4ojGEefLXry9Xsc2nCSwxpifjTGd6tKG6uDYnes8sx0i8pKI+Luc/5eIpIlItoisFZFr68iulsC9QIoxprkby5klIje7HjPGRBhjNrqrzIZEgKcNUMrlYmPMDE8b4aV0N8akO6L/E7AGeMc5lwtcDKwH+gDfi0i6MeYXN9vUGthrjMlyczmKG9GappchIsEickBEurgci3FqpbHO/i0iki4i+0RksogklJHXSTUSEbleROY472c7h5c5NbYrROQsEdnukj7ZyeOAiKwSkUtczo0VkTEi8o1To5svIu3Lua8/isgWEdkrIg+XODdWRJ5y2T/JjvIwxqQDc4EeLsceM8asNcYUGWPmAz8DA8qx7SIRWerc5y8i0s3l3GYRuU9ElovIQRGZICIhpeRxLjAdSHCe51jn+CXOszvgPMvkyuYtIsMduw6JyAYRGSoiTwO/A153bZ24thpEJFpEPhSR3c4z/7uI+DnnrheROU5tfL+IbBKRYZV51g0FFU0vwxhzFPgCGOVy+HLgJ2NMloicAzzjHIsHtgDjq1HOGc7b7k7TboLreREJBKYA04BYYDTwiYi4Nt9HAU8AjYF04OnSyhKRFOBN4I9AAtAUaFFVm8vIuzNWRNLLOB+KrW2uKuP8acD7wG2OXW8Bk0Uk2CXZ5cBQoC3QDbi+ZD5Oi2EYkOE8z+tFpCMwDvgzEAN8C0wRkaCK8haRvsCHwP1AI+AMYLMx5mHsn8DdTjl3l3JbrwHRQDvgTOBa4AaX8/2AdUAz4HngPRGR0p5PQ0RFs37zlVMDKd5ucY5/ysmieZVzDOBq4H1jzBJHYB8CBohIm1q2rT8QATxrjDlmjPkB+LqEXV8YYxYYYwqAT3Cp7ZVgBPC1MWa2Y/MjQFEN7VsiIrnYZvks4I0y0v0HWAZMLeP8LcBbxpj5xphCY8wHwFHs/Rfzb2NMhjFmH/aPpKz7LMkVwDfGmOnGmHzgX0AocHol8r4J+zlPd2rMO4wxaysq0OnbvQJ4yBiTbYzZDLyI/cMqZosx5h1jTCHwAfbPN66S9+TzqGjWby41xjRy2Yr75H4AQkWkn4i0xv6QvnTOJWBrlwAYY3KAvUBiLduWAGwzxriK25YS5biOEB/GimyZeRXvGGNysTbXhNOc8q7A1pzCSyYQkReALsDlpuzINa2Be13/vICWjs3FVPY+S1LysyrCPofKPMOWwIZKluNKMyDItVzK+dyMMYedt5W9J59HRdMLcX5cE7G1uquwtbRs53QG9ocOgIiEY5uVO0rJKhcIc9mvyohuBtCyuC/MoVUZ5VTETqwIACAiYViba2SnsUwE5gGPup4TkSewzeUhxphD5WSzDXi6xJ9XmDFmXGVsqICSn5Vgn0NlnuE2oKw+4vJCl+0B8l3LpfqfW4NERdN7+RRbi7qaE03z4uM3iEgPp9/tn8B8pxlWkqXA70UkzBkkuKnE+Uxsv1dpzMeK2QMiEigiZ2FHpKvcfwpMAi4SkUFOf96TnPzdXApcICJNRKQ5tg+wKjwL3Opci4g8hP2zOc8YU1GN9h3gdqdWLyISLiIXikhkFW0ojYnAhSIy2Okjvhfb9K/MKP572M95sIj4iUii038L5XxuTpN7IvC0iEQ6LZW/AB/X9GYaCiqa9ZspcvI8zeImOM6oby62ifedy/GZ2D7Bz7E1uPbAlWXk/zJwDPsj+wDb7+jK48AHTrP0ctcTxphjwCXY2toebJ/htZXpVyuJMWYVcBdW8HcC+wHX0fGPsP2Om7EDTxOoAsaYFdhpR/c7h/6JrV2luTzbv5Vx7SJsv+brjl3plDLQUx2MMeuAa7ADM3uwfzoXO8+2omsXYAdvXgYOYu+vuPb4KjDCGf3+dymXj8Z+dzYCc7DP/f2a3U3DQTQIsaIoSuXRmqaiKEoVUNFUFEWpAiqaiqIoVUBFU1EUpQqoaLqZkn7T9RERaeP4JtdZABcpI0SduPi/ezPio6HtFBVNpR7g/LFc72k73IWIPC4ij3vajpKosFcPFU2lytRljdRdOBPVG+T3v7Y+P1/4HlSHBvmlcSci0lNElogNhzYBcA3ldUrTs6xmqnNuloj8Q0TmOvlNE5FmLucrCit2v9iwYrki8p6IxInId05eM0SkcYkibxSRDBHZKSL3uuT1uIhMEpGPReQQcL3jhfKg2JBke0Vkoog0Kee53O/kmyEiN1biOZYamkxsWLP3nLx2iI2E7u9i58cuaU/qdnCe0dMiMhfrx92ugvw6iMhPYsOy7XE+z7LsLTO0XSlpSw3dV1o3iWPzzVK5kIAVhbD7q4gsB3JFZBx2gn+xA8UDUkrIPdfaaFW/ByIS4qTd69i0UES8P/CHMUa3Wto4EQjh/4BAbPSefOAp5/z1wJwS1xigQxn5zcIGZeiIjX4zCxtVCOdYLnCeU9YDWG+VIOf8ZuBXbHSaRCALWAL0BIKxQT8ec9K2cewYhw1s0RXYDZzrnH/cuY9LsX+0oVhXxl+xIdyCsSHTxpVxH0OxXkddnPw/Leu+nWeUj/XC8QfuwPpoFztifOWUFY4NSbcAuM3Fzo9d8iq+rwCX57kVSMUG4A6sIL9xwMPOPYcAg8q4vxQgBxueLRh4CSgofn4l0p6D9f45zUn7GjC7NHtdbL7Zef8+1g+++NxdwPfO+9Ocz7if89yuc74DwS7fh6VY3/ZQl2PnuuR3FrC9hL3H01T1e4ANpzcFGzfAH+gFRHn6d1rj37mnDfClzfnRHP+BO8d+oWai+XeX/TtdfiSPABNdzvlhgy6c5exvBq52Of858KbL/mjgK+d98Y+1s8v554H3nPePF/+wXc6vAQa77Mc7P6iAUu7jfRyxd/Y7lnXfzjNKd9kPc9I2x/4BHC3+0TvnRwE/uthZkWg+6XK+ovw+BN4GWlTwuT8KjHfZD8e6p5Ymmu8Bz7vsRzjPrU1Je11sLhbNc4GNLufmYl1XwcYj/UeJstYBZ7p8H24scX4zVRfNSn8PgBux3/9unv5t1ubWIPsk3EgCsMM43x6HLWUlriRlhQY7JayYiJQMK5bp8v5IKfslw31tc3m/BVvjLO0cWD/nL0XENTRcIVaISkbMSQAWl8i7PE4KTSY2/m0E0ARbO9wpJ2Li+pViW3m4pm1dQX4PAP8AFojIfuBFY0xpPtqnhLYTkbICgSRga/zFaXOctIlUHGnoeEhA7DNyDQnYGrhOREa7pA/i5BB2VXlOZVGV78FH2JrteBFphA0K8rCxsUO9FhXN2mUnkCgi4iKcrTgR9/CkEGfiRN2pJhm4iJpIlcKKlUVLoDjgRiunjGJKBinYhq25zK1EvieFfnPyrg7bsDXDZsYGNi5JZULIud5HufkZY3ZhuwkQkUHADBGZbewSGq7sBFz7k0uGtnOlvNB9uc7hMKA4XN3xe3D+GItDAmZyckjA4hB2pUbHL86igv2S309/bET58q6p6HvwBPCE2CDY32Jrv++VY2O9RweCapd52L6se0QkQER+D/R1Ob8MSBUbti0E29ypLjUJK1YWj4gNE5eKjaBTXjSh/2DDi7WG44MSw8ux9XoRSXEE5bHqGGeM2YmNcvSiiEQ5gxDtReRMJ8lS4AwRaSUi0dio9dXOT0RGikjxshv7sYJRWEpWFYW2c6XM0H3GmN1Y8bxGRPzFDpiVjJlZVkjA6oSwKxlCbj0Q4lwXCPwd209ZHmV+D0TkbBHp6ojvIWyzvbTn51WoaNYixob0+j22X24/9sv9hcv59dgf1AwgDRuWq7plVTusWDn8hB1Mmgn8yxgzrZy0rwKTgWkiko0dDOhXhq3fAa9gm5fpzmt1uRbb7FyNfcaTsP1oGGOmY4V+ObY74Oua5IddO2i+iORg7/VPxphNJTMwFYe2c01bUei+W7Ah7PZiB6x+KXF9WSEBqxPC7hng787I9n3GmIPYfvN3OVHzrWgBu/K+B82xz/MQtu/zJ3wgbqeGhlMURakCWtNUFEWpAiqaiqIoVUBFU1EUpQqoaCqKolQBr5un2axZM9OmTRtPm6Eoio+xePHiPcaYkvNST8HrRLNNmzYsWrTI02YoiuJjiEilvPe0ea4oilIFVDQVRVGqgIqmoihKFfC6Pk1FaQjk5+ezfft28vLyPG2KzxESEkKLFi0IDAys1vUqmopSD9m+fTuRkZG0adMGl7B1Sg0xxrB37162b99O27Ztq5WHNs8VpR6Sl5dH06ZNVTBrGRGhadOmNarBq2gqSj1FBdM91PS5ulU0RWSoiKwTu4jUg6WcbyUiP4rIb2IXALugojz35tQk8pmiKErNcJtoOoFHxwDDsAtPjRKRlBLJ/o5d56YnNqbgGxXluzvnaG2bqihKPeLAgQO88UaFUuAx3FnT7ItdIGujExh3PFAysrcBopz30Zy8vEKp5BcWkXu0tJUOFEWpjxQUVO33Wp5oFhZ6PvC7O0UzkZMXYdrOyYt+gV3u4RpnreVvsSsknoKI3Coii0RkEcCG3Tm1b62iKMfJzc3lwgsvpHv37nTp0oUJEyawcOFCTj/9dLp3707fvn3Jzs4mLy+PG264ga5du9KzZ09+/PFHAMaOHcvIkSO5+OKLGTJkCAAvvPACffr0oVu3bjz2WNkrnjz44INs2LCBHj16cP/99zNr1izOPvtsrrrqKrp2tctiffzxx/Tt25cePXpw2223HRfTadOmMWDAAE477TRGjhxJTk7ta4U7pxyV1ttaMkz8KGCsMeZFERkAfCQiXYwxRSddZMzb2KVUCY5PMmmZOXRr0cgtRitKfeOJKatYnXGo4oRVICUhiscuTi3z/Pfff09CQgLffPMNAAcPHqRnz55MmDCBPn36cOjQIUJDQ3n11VcBWLFiBWvXrmXIkCGsX78egHnz5rF8+XKaNGnCtGnTSEtLY8GCBRhjuOSSS5g9ezZnnHHGKWU/++yzrFy5kqVLlwIwa9YsFixYwMqVK2nbti1r1qxhwoQJzJ07l8DAQO68804++eQTLrjgAp566ilmzJhBeHg4zz33HC+99BKPPvporT47d4rmdk5egbAFpza/bwKGAhhj5jmLjTXDLnpfKgKkZWlNU1HcSdeuXbnvvvv461//ykUXXUSjRo2Ij4+nT58+AERF2V61OXPmMHq0bSB27tyZ1q1bHxfN8847jyZNmgC2Bjht2jR69uwJQE5ODmlpaaWKZmn07dv3+LzKmTNnsnjx4uO2HDlyhNjYWH799VdWr17NwIEDATh27BgDBgyojcdxEu4UzYVAkoi0xS7SdCVwVYk0W4HBwFgRSQZCgN3lZRoc4E9aZnZ5SRTFpyivRuguOnbsyOLFi/n222956KGHGDJkSKlTdcpbYyw8PPykdA899BC33XZbtewpmdd1113HM888c1KaKVOmcN555zFu3LhqlVFZ3Nan6awjfTcwFbsS3URjzCoReVJELnGS3QvcIiLLgHHA9aaCld6CA/20pqkobiYjI4OwsDCuueYa7rvvPn799VcyMjJYuHAhANnZ2RQUFHDGGWfwySefALB+/Xq2bt1Kp06dTsnv/PPP5/333z/ex7hjxw6yskpvUEZGRpKdXXbFaPDgwUyaNOn49fv27WPLli3079+fuXPnkp5ul6U/fPjw8VpvbeJWN0pjzLfYAR7XY4+6vF8NDKxKniEB/mzbf5gjxwoJDfKvHUMVRTmJFStWcP/99+Pn50dgYCBvvvkmxhhGjx7NkSNHCA0NZcaMGdx5553cfvvtdO3alYCAAMaOHUtw8KlLpQ8ZMoQ1a9Ycby5HRETw8ccfExsbe0rapk2bMnDgQLp06cKwYcO48MILTzqfkpLCU089xZAhQygqKiIwMJAxY8bQv39/xo4dy6hRozh61E5NfOqpp+jYsWOtPhuvW8I3KbW7yb/4n3w9ehBdEqM9bY6iuIU1a9aQnJzsaTN8ltKer4gsNsb0ruhar3OjDAmwtct0baIriuIBvC7KUVCgHwF+wnodDFIUr2bv3r0MHjz4lOMzZ86kadOmHrCocnidaArQtlm4DgYpipfTtGnT43MxvQmva54DJMVFaPNcURSP4JWi2SE2ki17c8nL97wfqqIoDQuvFM2k2AiKDGzak+tpUxRFaWB4pWh2jIsE0MEgRVHqHK8UzTbNwvD3E+3XVBSlzvFK0QwO8Kd10zDSMlU0FcUXiYiIAKw754gRIzxszcl4pWiC7ddMy9LmuaLUd6oahNiVhIQEJk2aVIvW1Byvm6dZTFJsJDPWZHGsoIigAK/VfkWpmO8ehF0rajfP5l1h2LNlns7NzeXyyy9n+/btFBYW8sgjj9CuXTv+9Kc/kZubS3BwMDNnziQwMJA77riDRYsWERAQwEsvvcTZZ5/N2LFj+eabb8jLyyM3N5cffviBF154gYkTJ3L06FEuu+wynnjiiQrN3Lx5MxdddBErV65k7NixTJ48mcOHD7NhwwYuu+wynn/++dp8KpXCe0UzLoLCIsOmPbl0ah7paXMUxafwZBDi8li6dCm//fYbwcHBdOrUidGjR9OyZcuKL6xFvFc0Y61QpmVlq2gqvk05NUJ3Ud+CEBczePBgoqNtoJ6UlBS2bNmiollZ2sWE4yfoYJCiuIH6FoS4GNewc/7+/jXqL60uXtsZGBLoT6smYTrtSFHcgCeDENd3vLamCdadUkfQFaX28WQQ4vqO1wUh7t27t1m0aBEAz32/lnd/3sjqJ4cS6O+1lWZFOQUNQuxeGlQQYlc6xkWQX2jYsld90BVFqRu8unl+fAQ9M4cOsTqCrijehAYh9gDtYyIQseugD/O0MYpSyxhjSh2x9hU8FYS4pl2SXt08Dw3yp0XjUI3irvgcISEh7N27t8Y/cOVkjDHs3buXkJCQaufh1TVNsE30NA0Rp/gYLVq0YPv27ezevdvTpvgcISEhtGjRotrX+4BoRjAnbQ8FhUUE6Ah6g2PljoM8/c0a7j6nAwM7NPO0ObVGYGAgbdu29UjZqzMO8fzUtRQWGYID/AgO8Cc40HkN8Dv5fYAfIYH+RIYEcEHX+AYxi8X7RTMukmOFRWzdd5h2MRGeNkepQ9Kzcrj2/QXsyz3Ggs37eOTCZK47vY1P9wO6m10H87hx7EKOFRbRumkYe3OKOFpQyNGCIrvlF5JXUMSxgqJTrs09WshV/Vp5wOq6xftFM9YKZVpWjopmA2LbvsNc8+58/ET4evQgXpmRxuNTVrN2VzZPDu+ika+qweFjBdz84UKy8/KZdMfpJMdHlZm2qMhwrNAR0oJCrn5nPpMWb2sQoun136z2jmiqOyXkHi3g5enreXn6elZnHPLZQYSs7Dz++N58Dh8r4KOb+tIlMZq3/9iLu8/uwPiF27j63V/Zk3PU02Z6FUVFhj+PX8rqjEO8dlXPcgUTwM9PCAn0Jzo0kNjIEEb2bsGSrQfYsNv3f4deL5oRwQEkNgpt8INB8zbsZeirs/n3D2n8+4c0Lvj3z5z5wiye/mY1izbvo6jINwT04OF8rn1vAZmHjvLfG/oe/3H7+Qn3nd+J10b1ZMWOg1zy2hxWZRz0sLXew3NT1zJtdSZ/vzCFczrHVfn6S3sk4u8nfL54uxusq194vWgCdIiNYH0DjXZ0+FgBj09exah3fsVPhAm3DmDB387lmd93pV1MOGN/2cyI/8yj3zMz+duXK/hp/e5S+6O8gdyjBdwwdgEbd+fy9rW96NW68SlpLu6ewGe3nY4BRrw5j2+W76x7Q72MiQu38dZPG7m6XytuGNimWnnERoVwZscYvliyg0If+YMuC6/v0wTrTvnrxr0UFhn8/RrOIMDCzfu4/7NlbN57mOtPb8MDQzsRFmQ/0lF9WzGqbysO5eXz49ospq3K5KvfdvDp/K1EhgQwuHMs56c258xOMcevqc8cLSjkto8Ws3TbAd64+jR+lxRTZtquLaL5390Duf2jxdz16RLWZSbx58FJ+DWg70ZlmbdhL3/7cgW/S2rG45ek1mgQbUSvFty5dglz0/dwRseyPx9vp/7/WipBUmwkRwuK2L7/MK2bhld8gZeTl1/Iv6au4725m0hsFMq4W/ozoH3pbmdRIYEM75HI8B6J5OUXMidtD1NX7WL6mky+WppBcIAf/ds1pXfrxvRq05geLRvVOxEtKCzinnG/MSd9Dy+M6MbQLvEVXhMbGcK4W/vz8Jcr+ffMNNbtOsRLl/cgPLh+3Zsn2bg7h9s/XkybZuG8ftVpNZ4uNDg5lujQQD5bvF1Fs77TIc4ZQc/M8XnRXLJ1P/dNXMbGPblc078VDw1LrrQQhAT6c25KHOemxFFQWMSCzfuYtiqTXzbs4cXpdhJ1gJ+QmhBFr9ZN6N2mMb1bNyY2qvreEzWlqMjw189XMHVVJo9elMLI3pWP0h0c4M8LI7qRHB/F09+s5g9v/sI71/amZZMwN1rsPowxrNxxiO9X7SQvv4gbBrahRePq3cuBw8e46YNF+PsJ71/Xh+jQwBrbFxzgz/AeCYxfuI2DR/JrJc/6iG+Ipsu0o3NTqt6J7Q3k5Rfy8oz1vDN7I/HRoXx8Uz8GJVV/MneAvx+nt2/G6e1tHgcOH2PJ1v0s2ryfRVv288n8Lbw/dxMArZqEHa+J9mnThA4xEXXS1DXG8OTXq/l8yXb+79yO3Dio6pO9RYSbBrUlKTaCuz9dwiWvz+GNq3uVWTOvbxQWGRZt3sf3q3YxbVUmOw4cwd9P8Bfhw3mbGdW3FXed3YG4KvyxHSso4raPFrNj/xE+vaUfrZrW3p/IyF4t+XDeFr5ensHV/VrXWr71Ca+Op+lK/3/O5PQOTXnp8h4esMq9LNt2gHs/W0Z6Vg6j+rbkbxckExni3n/xYwVFrMo46IjoPhZv2c+enGMAtGwSykPDkhnWpblbJ5K/PH09r85M48aBbXnkouQal7VpTy43f7CQLXsP07NVIzrERpIUG0FSXARJsZHERQXXi4nxRwsK+WXDXqau3MX01ZnszT1GUIAfZyQ14/zU5pybHMeR/EJe/zGdiQu34e8nXDugNbef2Z6mEacGAHbFGMMDk5bz2eLtvHJFDy7tmVirthtjGPrKz4QF+/PlnQNrNW93U9l4mj4jmn98bz4HDuczZfQgD1hVu+TlF7JpTy4bduewaPN+Pvp1CzERwTw3ohtneqivyBjDlr2HWbB5H+/P2cTaXdn0bduERy9KoUtidK2X996cTfzj69WM7NWC5/7QrdZqtofy8nl1RhrLtx8gLSuHA4fzj5+LDA6gfWzESULaITaCxEahJ5VvjKGwyJBfaCd45ztbgcu+nwjBAX4EBZzscliWq2/u0QJ+Wr+bqat28cOaLLKPFhAe5M85yXEMdQbsIkrphtm69zCvzkzjy9+2ExLoz40D23LL79oRHVb6n+p/ftrAs9+t5Z5zOvCXIacuS1EbvDN7I09/u4YZfznzeCvQG2hwovnklNWMW7CVVU+c7xWjpMYYduccZUNWLhv35LAhy4rkht057DhwhOKPRQRGnNaCv1+UUm/6iAqLDOMXbuXFaevZf/gYI3u14L7zOxEbWTt9nxMXbeOBScsZ1qU5r43q6baYAsYY9uYeIy0zh/SsbNKyckjLzCEtK+ekyfEhgda/2lUUq/uz8RNcfLlPiOq2fYc5WlBE47BAzkuJY2iX5pzevhkhgf6Vyjc9K4dXZqzn6+U7iQwJ4NbfteOGQW1PEtrvV+7ijk8Wc2HXeF4b1dNtteqs7DwGPPMDt57Rjr8O7eyWMtxBgxPNcQu28tAXK/j5gbPrbUf/pj25vD17I2t2HmLD7hyy806spBca6E+7mHDax0Qcf20fE0HbZuGEBlXuh1PXHDySz+s/pDH2l80EB/hz19kduGFgm0r/0F3JPVrArHW2pvX18gwGdmjGu9f1JjjAM/e+P/cY6btzSM+yW3FAmEB/P4L8hUB/PwIDTt635+17Yzjhs51faF0O80+4HR4r9uV29mMjQzg/tTl92jSu0Z/Emp2HeGn6eqavzqRxWCB3nNWeP/ZvQ3pWDiPf+oXk+CjG3dK/Wp9RVbhp7EJWZhzklwcHe800wAYnmos272PEf+bx3+v7cHbn+rVYU3ZePq//kM77czcR6O9Hj5aNHFEMp31sBO1iIoiPCvGKGnJpbNqTy9PfrGHGmkxaNgnl4QuSOT+14v7OvTlHmbEmk6mrMpmTvodjBUU0CQ9iWJfmPHxhcr2b+uRNLN12gJemr2f2+t3ERAZjDAQH+PHVXQOJiSy/37M2+G7FTu74ZAkf3NjXY11KVaWyoukz38oTI+jZ9UY0i4oMkxZv5/mpa9mTY5ux9w+tvWZsfaFts3Deva43c9L28I+vV3P7x0vo17YJj16cQmrCyf2d2/YdZqozErxoyz6KDCQ2CuWafq05PzWOXq1rVtNSLD1aNuLDG/uyYNM+/jVtHet2ZfPJzf3qRDABzkmOpVFYIJ8t2uY1ollZ3CqaIjIUeBXwB941xjxbSprLgccBAywzxlxVnbIahQURExlcb9wpF27exxNTVrFyxyFOa9WI967rQ/eWjTxtllsZlNSMb+4ZxLiF23hp2jouem0OV/RuyYheLZiTvoepqzJZs/MQAJ2bR3L3OUmcnxpHSnxUvRi19kX6tm3CxNsG1Lm3XHCAP8O7JzBu4TYOHs4vc2DKG3GbaIqIPzAGOA/YDiwUkcnGmNUuaZKAh4CBxpj9IlK+BLgZAAAgAElEQVSjKmLHuAiPL32x48ARnv1uLVOWZdA8KoRXr+zBJd0TGowoBPj78cf+rbmkewKvzbT9neMXbkMEerduzMMXJDMkNc7nnRDqG57oVxzZuyUfzNvClOUZXNPfd+ZsurOm2RdIN8ZsBBCR8cBwYLVLmluAMcaY/QDGmKyaFJgUG8lni7Z5ZEGqI8cKeWv2Bv7z0waMgXsGJ3H7me0abL9cdGggf78ohav7t2b59gOc3r5ZnTUNlfpBakIUnZtHMmnxdhXNSpIIbHPZ3w70K5GmI4CIzMU24R83xnxfMiMRuRW4FaBVq7KDnHaIjSD3WCE7D+aR0Ci0ZtZXEmMMXy/fyTPfriHjYB4XdovnoWGdq+3e5mu0bRZO22Zaq2yIiAgjerXgqW/WkJ6V7TPLbLuzx720ql7JofoAIAk4CxgFvCsip3T8GWPeNsb0Nsb0jokpu1PZNYp7XbBt32GueOtXRo/7jUZhQUy4tT9jrjpNBVNRHIY7cTYnLd7haVNqDXeK5nbANbpCCyCjlDT/M8bkG2M2AeuwIlotkuLsP1ldBCRetHkfw8fMZe2uQzzz+65MGT2Ifu28w59ZUeqKmMhgzu4Uw5e/bfeZOJvuFM2FQJKItBWRIOBKYHKJNF8BZwOISDNsc31jdQtsEh5E0/Ag0tw8gv7Fku1c9c58okMD+equgYzq28prJvAqSl0zolcLMg8dZXaabyxH7DbRNMYUAHcDU4E1wERjzCoReVJELnGSTQX2ishq4EfgfmPM3pqUmxQXQVqWe2qaRUWGF6au5S8Tl9G7TWO+vPN0XcxNUSrgnM5xNA4LZJKPLIXh1qFdY8y3wLcljj3q8t4Af3G2WiEpNpKvlu6o9RH0w8cKuHfiMr5buYtRfVvx5PDUBrHGs6LUlKAAP4b3SOTT+Vt9Ys6mz/3qk+IiyM4rICu79lYj3HUwj8vfmsfUVbt45KIU/nlZFxVMRakCI3q14FhhEZOXlxzW8D587pdf7E65vpYGg1ZsP8jwMXPYtDuXd6/rzU2D2jaYieqKUlu4ztn0dnxONJNii0fQaz4Y9P3KnYx86xcC/Pz4/M7Tq7W0qaIoJ+ZsLtt2wOuX2/Y50WwWEUTjsMAazdU0xjDmx3Ru/3gJKfFRfHXXQDo3j6pFKxWl4XFpz0QC/IRJS7y7tulzoikiJMVGkl7NEfSjBYXcO3EZL0xdx/AeCXx6S391/1OUWqBZRDBndYrlyyU7KCgs8rQ51cbnRBPs6pTrM3OoaqzQfbnHuPqd+Xzx2w7uPa8jr1zRw+3BWhWlITGiVwuyso/yc9oeT5tSbXxSNJNiIzh4JP/4QmCVoXiEfMWOg4y56jRGD07SAR9FqWXO6Rzr9XM2fVQ0q+ZOuW3fYS5/ax47Dxzhgxv7cmG3eHeapygNluI5m9NXZ3LgcOUrNfUJ3xTNuMoH7kjPymHkf+Zx8Eg+n9zSn/7qP64obmVkbztn8w9v/sJrM9PYsjfX0yZVCZ8UzdjIYKJCAip0p1ydcYgr3ppHQVER42/tTw8fj6yuKPWB1IRoXrmiB03Cg3hx+nrOfGEWw8fM5b05m8g8lOdp8yrEJyPkighJcZHlztX8bet+rnt/AeHBAXxycz/1IVeUOuTSnolc2jORHQeO8PWyDCYvy+AfX6/mqW9W079tUy7pkcCwLs1pFBbkaVNPwWdWoyzJg58vZ/rqTBY/ct4p5+Zt2MvNHyykWWQwH9/Ur94u+asoDYn0rBymLMtgyrIMNu7JJdBfOCMphkt6JHBuchzhwe6t4zW41ShL0iE2gvELt7E35yhNI07Ms/xxbRa3f7yYVk3C+PjmfsRF+dbKkIrirXSIjeD/zuvIn89NYlXGISY7AjpzbRahgf68emUPhqQ297SZvtmnCS4BiV0Gg75bsZNbP1pEUlwEE24boIKpKPUQEaFLYjR/uyCZuX89hwm39ic6tP5MU/JZ0exYYgR90uLt3PXpErq1aMSnt/SnSXj96ytRFOVk/PyEfu2a0qtNY9bsOuRpcwAfFs3mUSFEBAeQnpnNR/M2c99nyxjQvikf3dSXqBDvjuenKA2NlPgotu07QnZevqdN8d0+TRGhQ2wEX/62g0N5BZybHMfrV/VUt0hF8UKS421329pd2fRp08SjtvhsTROsO+WhvAIu7p7Am9ecpoKpKF5KcryNMrZmp+eb6D5b0wS4fmAbOsRGcPPv2unCZ4rixTSPCqFRWKCKprtJTYgmNSHa02YoilJDRITk5lGs3un5AMY+3TxXFMV3SI6PYt2uQx5fP11FU1EUryA5PpK8/CI2ezjAh4qmoiheQX0ZDFLRVBTFK0iKiyDAT1Q0FUVRKkNwgD/tYyJY4+HBIBVNRVG8huT4SK1pKoqiVJbk+Ch2Hszz6FIZKpqKongNxYNBqz1Y21TRVBTFazgxgu65fk0VTUVRvIaYyGCaRQR7tF9TRVNRFK/C04NBKpqKongVKfFRpGXmUFBY5JHyVTQVRfEqkuOjOFZYxMY9nnGnVNFUFMWr8LQ7pYqmoiheRbuYcIL8/Tw27UhFU1EUryLQ348OsZ5zp1TRVBTF60iOj9LmuaIoSmVJjo9kd/ZR9uQcrfOyVTQVRfE6Ujw4GORW0RSRoSKyTkTSReTBctKNEBEjIr3daY+iKL6BJ0fQKxRNEYkSkfalHO9WwXX+wBhgGJACjBKRlFLSRQL3APMra7SiKA2bxuFBNI8K8chgULmiKSKXA2uBz0VklYj0cTk9toK8+wLpxpiNxphjwHhgeCnp/gE8D+RV2mpFURo8nnKnrKim+TeglzGmB3AD8JGI/N45V9FC4onANpf97c6x44hIT6ClMebr8jISkVtFZJGILNq9e3cFxSqK0hBIjo8iPSuHowWFdVpuRaLpb4zZCWCMWQCcDTwsIvcAFa2jWZqoHr9GRPyAl4F7KzLSGPO2Maa3MaZ3TExMRckVRWkAJMdHUVBkSM/KqdNyKxLNbNf+TEdAz8I2s1MruHY70NJlvwWQ4bIfCXQBZonIZqA/MFkHgxRFqQyeiq0ZUMH5OyghrMaYbBEZClxewbULgSQRaQvsAK4ErnLJ5yDQrHhfRGYB9xljFlXaekVRGixtm4UTEuhX5/2a5dY0jTHLjDFppZyqMCaTMaYAuBuYCqwBJhpjVonIkyJySbWsrW8U5sPW+XB4n6ctUZQGh7+f0Cmu7geDyq1pikgUcBd2AGcyMB0rhPcBS4FPyrveGPMt8G2JY4+WkfasyhpdL9g8F769D7JW2/2YZGg9AFqdbl+jW3jWPkVpACTHRzF11S6MMYhUNDZdO1TUPP8I2A/MA24G7geCgOHGmKVutq1+kp0J0x+B5RMguhUMHwM5mbBlHiz/DBa9b9M1auUIqLM17QB19KEqSkMhOT6K8Qu3kXnoKM2jQ+qkzIpEs50xpiuAiLwL7AFaGWM8u1q7JygsgIXvwI//hII8OON+GPQXCAqz538HFBVC5krY8ovd0mfA8vH2fHgMtBoAbQZBj6sgONJjt6IovoKrZ1B9Ec384jfGmEIR2dQgBXPLPNsUz1wJ7QfDBS9A01OcpMDPH+K7263/HWAM7E2HLXNtHlt+gTWTYf5/4A/vQmKvur8XRfEhOsfbysfqnYc4u3NsnZRZkWh2F5HiXlYBQp19AYwxJsqt1nmanCyY/hgs+xSiWsDlH0HyxZVvZotAsyS79breHts8B764Dd4bAuc8AqffA34aN0VRqkNUSCAtGofW6WBQuaJpjPGvK0PqFYUFtm/yh6cg/7Bthp9xHwSF1zzvNoPgjjkw+R6Y8Rhs/BEuewsim9c8b0VpgNR1bE2t4pRk63x45yz47n5I7Al3zoNzH6sdwSwmtDFc/iFc/Kot783TYd33tZe/ojQgkuOj2LQnl7z8unGnVNF0ZdVX8P4QyN0LI8fCH7+yTWt3IGKb7Lf9BJEJMO4K+PYByNe4JYpSFVLiIykysD6zboZbVDSLMQZmvwAxneHuhZB6Wd1MEYrpBDfPgH63w4K34N3BsHud+8tVFB+hrmNrqmgWs+knOzo+4G4IjqjbsgNDYNhzcNVEyN4Jb50Ji8daIVcUpVxaNg4jPMi/znzQVTSLmTfGzqXsOtJzNnQ8H+74BVr1gyl/gonXwpH9nrNHUbwAPz+hU/PIOlvSV0UTbHM4bRr0ucXW+jxJZHO45ks470lY9y28OcgOEmmtU1HKpHgE3dTB70RFE+DXN8A/GPrc5GlLLH5+MPBPcNN0CAy1g0T/HWZH2hVFOYXk+Ciy8wrYceCI28tS0czdC8vGQ/crIbxZxenrksTT7JSni16GfRvtyP64UZC1xtOWKUq9oi5ja6poLnrf+pL3v9PTlpSOfyD0vhHu+c16EG2eY+d1fnUnHNhW8fWK0gDo3DwSkboZQW/YollwFBa8DR3Og9jOnramfILCrVfSn5ZZgV8xCV7rBVMf1nieSoMnPDiA1k3CVDTdzopJkJsFA+7ytCWVJ6wJnP80jF4MXUfY/thXu9s5psdyPW2doniMunKnbLiiaYydZhSbCu3O8rQ1VadRS7j0DTtFqc0g6yf/756w8D0bUV5RGhjJ8VFs2XeY3KMFbi2n4YrmxlmQtcrWMr05OHBsMowaBzdOhSbt4Ju/wJsDYcOPnrZMUeqU5PgojIG1u9w7GNRwRXPeGAiPtU1cX6BVf7jhO7jyUyg8Ch9dChOugf1bPG2ZotQJyU5sTXc30RumaGathfTp0PdWCAj2tDW1hwh0vhDunA/n/B3SZsCYvjDrWch3//w1RfEkiY1CiQoJUNF0C7++AQEhdiqPLxIYYpfjGL0IOg2DWc9Y8VwzRT2LFJ9FROhcB4NBDU80c/c4k9lHQXhTT1vjXqJb2BB3102BoAjbXP/oMti93tOWKYpbSImPYu2ubIqK3Fc5aHiiufA92+dXXyezu4O2Z8BtP8PQ52DHEnhzgJ3fmVe360UrirtJjo/k8LFCtu477LYyGpZo5ufZFSWTzoeYjp62pm7xD4D+t9v5nd1H2YGw13vD0nFQVORp6xSlVqiL2JoNSzRXfAa5u71rMnttExEDw1+HW2ba5vtXt8P750PGb562TFFqTMe4SPzc7E7ZcESzeDJ7XFfbXG3oJPaCm2bA8DGwfxO8fTZMHg05uz1tmaJUm5BAf9rFRLDajYE7Go5obvgBdq+BAXd692T22sTPD3peY5vsA+6CpZ9af/Z5b6hXkeK1uNudsuGI5rwxEBEHXf7gaUvqHyHR1p/9jnnQojdMfcjxKvrB05YpSpVJjo9kx4EjHDzinj/+hiGamathw0zoe4tvTWavbWI6wjWfw6jxjlfRZTD+ati3ydOWKUqlKR4MWuum2mbDEM1f34CAUOjlo5PZaxMROyH+zvkw+FHrwz6mH8z8h0ZRUryClGLRdJMPuu+LZs5uWD4RejSAyey1SWAI/O5e61WUMhx+/he83seG01OvIqUeExsZTOOwQFZnaE2zeixqgJPZa5OoBPjDOzaKUlhT+PwmeO88WD0Zigo9bZ2inIKIkJIQ5bbVKX1bNPPzYME70HEoNEvytDXeTav+cOssuPhVyMmEiX+08Tt/fROO1s1604pSWbokRLNuVzb5hbXvuOHborl8Ahze07Ans9cmfv7Q63q4Zylc/iFExsP3D8JLKdYtU9csUuoJKQlRHCssIj0rp9bz9l3RLCyAOS9DfHdo8ztPW+Nb+Pnbfs6bpsLNM6HDubbG+Wp3+OwG2L7Y0xYqDZwuidEArNxxsNbz9l3RXPm59XQ54wGdzO5OWvSGkf+1C74NuBPSZ8C758B7Q2D1/7TfU/EIbZuGExbkzyo3DAb5pmgWFdrR3thU6HSBp61pGDRqCUOegr+sttGUsnfBxGtP9HvqdCWlDvHzE5Ljo1iVoTXNyrH6f7BnvV3y1s83b7HeEhxpoynd8xtc8fGJfs+Xu8Cs53S5YaXO6JIQxeqMQ7UeW9P3FKWoCGb/C5p1tP1uimfw84fki22/503ToWU/mPVPeKWrHTQ6tNPTFio+TmpCNLnHCtlSy7E13SqaIjJURNaJSLqIPFjK+b+IyGoRWS4iM0WkdY0LXfetXWXyd/fZH67ieVr2havG2+WGOw1z1mrvBpPvgb0bPG2d4qOkJFjPoNoeDHKbaIqIPzAGGAakAKNEJKVEst+A3saYbsAk4PkaFWoMzH4eGrfVwBz1kbhU+MO7MHqJja60bLwNhDzpRti1wtPWKT5Gx7hIAv2l1geD3FnT7AukG2M2GmOOAeOBk9rLxpgfjTHFdedfgRY1KjFtOuxcZt3//ANqlJXiRpq0hYtehj8vh9NHw/pp8J9B8MlI2DLP09YpPkJQgB8d4yJrfTDIncqSCLjOdt4O9Csn/U3Ad9UurbiWGd0Kul9Z7WyUOiSyOZz3JAz6P1j4rh1l/+9QGyA5sRc0TbKeXM2SICpRp44pVaZLQjTT12RijEFq6fvjTtEszcJSh7FE5BqgN3BmGedvBW4FaNWqVemlbZwF2xfChS+Bf2A1zFU8Rmhju+Rw/7vgt4+sJ9ey8XDUpVkVGAZNOzgi2vHE+6YdICjcc7Yr9ZrUxCgmLNrGrkN5xEeH1kqe7hTN7UBLl/0WQEbJRCJyLvAwcKYx5mhpGRlj3gbeBujdu3fp8wdmvwCRCbavTPFOgsKg3212M8b6uO9Js9PH9qbb99sXwcovOOn/t1Fru4Z931tUQJWTSD0+GHTIK0RzIZAkIm2BHcCVwFWuCUSkJ/AWMNQYk1XtkjbPhS1z7aRqDTLsG4jY5ntkc2hbwg02Pw/2bbAiujcNNs+BGY/BvNdh4J+hz00QWDs/EMW7SY6PQgRWZRzkvJS4WsnTbaJpjCkQkbuBqYA/8L4xZpWIPAksMsZMBl4AIoDPnP6GrcaYS6pc2OznITwWel1Xezeg1F8CQ+xIfFyq3T/jftg6384DnfYw/PJvOxh42nU2rdJgCQsKoF2zcFbuqL0RdLcOMRtjvgW+LXHsUZf359a4kG0LbX/mef/Q2kVDplU/uPZ/ttb54z/huwdg7qtWPHv+EQKCPG2h4iG6JEazcFPteaJ5v0fQ7OchtInt01KUNoPg+m/g2sl2Xfdv/gKvnQaLP9AVNhsoqQlRZBzMY1/usVrJz7tFM+M3SJtm42UGR3jaGqW+IALtzrTR5q/5AiJiYco9diL90k9t2EClwZCaYMPE1dZ8Te8Wzdn/ssvP9r3V05Yo9RER6DDYxvy8aqL9rnx1B4zpa787mat0vaMGQPEIem15Bnmv28yulbD2azjzQQiJ8rQ1Sn1GBDqeD0lDbGyCOS/DD/+wW3Qr6DTULonSZpDOvvBBGoUFkdgotNZ80L1XNH/+FwQ5YcgUpTKIQOcL7XZoJ6RNhXXfw5KPYMHbEBQB7c+xQUWShkB4M09brNQSXRKjam11Su8Uzd3rYNVX1v0utLGnrVG8kah4u95Rr+vh2GHYNNvWQtdPhTWTAbHRmToOtSIa01ndOL2Y1IRopq3OJOdoARHBNZM97xTNn1+004t0wTSlNggKs030TkNtPNadS2H997DuO5j5hN1iU+G0P0K3KyCsiactVqpIakIUxsCanYfo06Zmn5/3DQQVHIUVn9kpRtp8UmobPz9IPA3O/hvc/jP832q44F92nuf3D8KLnWwou42zrMAqXkHxQmuraqFf0/tqmjmZ4B8Ep9/jaUuUhkB0ovVp73uLjfm55CNYPt4u3NeotZ043+Mqm06pt8RGBtMsIoiVtdCv6X01zSP7rHtcZO34kSpKpWneFS54Hu5dB79/Fxq3hh+fgle6wCeXw5opOoG+niIipCRE18q0I++raSIw8E+eNkJpyASGQreRdtu3EX77BJZ+AhOusTEQeoyC9oOtqEa10IDY9YQuCVG8PXsjRwsKCQ6o/lI43vdpRsRpU0ipPzRpB4MfgbMesmu+//YR/PK69XsHEH/7fW3U2m6NW0OjVifeRzTXFVPriNSEaAqKDOt35dC1RXS18/E+0Yxs7mkLFOVU/ANOjMDn7Ias1XBgC+zfAge22vfpMyBnV4nrgiC6JcR3hy6/hw7naWQmN9Elsdgz6GADE01Fqe9ExEBEqYsQ2FigB7edLKj7N9t5oqu+sA4byRfZhQHbnaWrENQiLRuHERkcwMoa+qCraCpKXRIYcmLdI1cKC2DzbDsqv2YKLBtnHTeSL7EC2maQLkldQ/z8hOSEqBoPBqloKkp9wD/AunC2PwcufBk2/GAFdOXnsOQD25efcqkV0BZ9tB+0mnRJiObTBVsoLDL4+1XPw0tFU1HqGwFBJ/pHjx224Q+LxXPBW3ZEvtMwSOgBzbtZF08NslwpUhOiyMsvYuPuHJLiIquVh4qmotRngsIg9VK7Hc22rp0rJtm4oAvfsWn8Aq1wNu/qsnXRuAylcNwzKOOQiqai+DzBkdDtcrsVFcK+TbBrufVU2rUCNsyEZZ+eSB/d6mQRbdYJmrRt0INL7WPCCQ7wY1XGQS7tWb2piyqaiuKN+PlDsw526/L7E8ezMyFzxQkh3bncRm8qXvLYL8DOLW3W0W4xnU6sJR9cvZqXNxHg70fn5pE1WmhNRVNRfInIOLt1cFmz8FguZK2168fvWWeXPt69zjb1TeGJdFGJjoB2gpiOENcFYlN8Lsh3amI0Xy/LwBiDVCPcn4qmovg6QeHQopfdXCk4Bvs3WTHd7YjpnnXWJfRYzol0jVrbJn5cqhXS5l2gURuvHcFPTYji0/lb2b7/CC2bhFX5ehVNRWmoBATZ5nlMJ0i++MRxY+DgdruGUuYK+7prJaz9huPN/KAIWwtt3sUR0m6Q2MsrhNR1oTUVTUVRao4INGppt05DTxw/dhh2r7ECmrnSiumKz2HR+/Z847bQ52boeXW9Hrnv3DwSfz9hVcYhhnaJr/L1KpqKolSOoDBbm0x0aeYbY91Ct/5qxXPaw/DDUzYCVN9bbbO+nhES6E+HmIhqL7SmoqkoSvURcaI2tbJToXatgAXvwPKJsORDaDXABnDufHG9moCfmhjFnLQ91bq2/ndAKIriPTTvCpf8G+5dA0OehuyddnmQV7rAj8/YVUDrAakJ0WRlHyUrO6/K16poKopS+4Q2htPvhtG/wdWTbOi7n56z4vnZ9bB5LhzeZ/tJPbDWUmpCcZi4qs/X1Oa5oijuw88Pks6z294Ntt/zt49g1Zcnp/MPgoAQCAiGgFDnNcRGhQpwtuSL4bRra2Up5RRHNFdnHOLsTrFVulZFU1GUuqFpezj/aTj7YeullLsbCvLsCrP5R+xr8X6By35+no05OuUeWDkJLv63dQetAVEhgbRuGlatwSAVTUVR6pagMOg6omrXGAOLx8K0R+DN0+Hcx6HPLTWaF9olIZoV1RBN7dNUFKX+IwK9b4C7foXWA+G7B+C/w6wXUzVJSYhi677DHDxStRVEVTQVRfEeolvA1Z/BZW/B7rXw5kCY84qNfF9FUl36NauCiqaiKN6FCHS/Eu5aYAeYZjwG751rPZSqgKs7ZVVQ0VQUxTuJjIMrPoaRY+HANnjrTJj1nA1EUgliIoOJiwqu8rQjFU1FUbwXEUi9zNY6Uy+FWf+Ed86GjN8qdXlqQrTWNBVFaYCEN4U/vAtXjoPcPfDOYPjuQbsGfTl0SYgiPSuHI8cKy03nioqmoii+Q+cL4K750PMauwjdq91hxuPW+6gUUhKiKTKwdlflm+gqmoqi+Bahjaz/+10LbGi7Oa9Y8fzxGcg7uSneJbHq7pQqmoqi+CbNkmDE+3DHL9DuTPjpWXilG/z8Ihy1kekTG4USHRpYpX5Nt4qmiAwVkXUiki4iD5ZyPlhEJjjn54tIG3faoyhKAyQuxY6y3/oTtOwHM5+0Nc9fXkcK8khNiKofNU0R8QfGAMOAFGCUiKSUSHYTsN8Y0wF4GXjOXfYoitLASegBV0+Em2bYZTqmPQyv9uC6gGls2Fl6n2dpuLOm2RdIN8ZsNMYcA8YDw0ukGQ584LyfBAyW6iwPpyiKUlla9oFr/wfXfwNN2nH+lheZFvDnSl/uTtFMBLa57G93jpWaxhhTABwEmpbMSERuFZFFIrJo9+7ypxAoiqJUijaD4IZv2XHxOLJM5dc0cqdollZjNNVIgzHmbWNMb2NM75iYmFoxTlEUBRGa9xzGVeapSl/iTtHcDrR02W8BZJSVRkQCgGig8p0LiqIoNcTfT0iOj6p0eneK5kIgSUTaikgQcCUwuUSaycB1zvsRwA/GmFNqmoqiKO5kWBWW8nWbaDp9lHcDU4E1wERjzCoReVJELnGSvQc0FZF04C/AKdOSFEVR3M0tZ7SrdFq3Rm43xnwLfFvi2KMu7/OAke60QVEUpTZRjyBFUZQqoKKpKIpSBVQ0FUVRqoCKpqIoShVQ0VQURakCKpqKoihVQEVTURSlCoi3OeCIyG5gi6ftqCHNgD2eNqIOaWj3C3rP3khrY0yFwS28TjR9ARFZZIzp7Wk76oqGdr+g9+zLaPNcURSlCqhoKoqiVAEVTc/wtqcNqGMa2v2C3rPPon2aiqIoVUBrmoqiKFVARVNRFKUKqGjWISKyWURWiMhSEVnkaXvcgYi8LyJZIrLS5VgTEZkuImnOa+VXsfICyrjnx0Vkh/NZLxWRCzxpY20iIi1F5EcRWSMiq0TkT85xn/6ci1HRrHvONsb08OH5bGOBoSWOPQjMNMYkATPxvQj9Yzn1ngFedj7rHk5Abl+hALjXGJMM9AfuEpEUfP9zBlQ0lVrGGDObUxfHc13f/gPg0jo1ys2Ucc8+izFmpzFmifM+G7ucTSI+/jkXo6JZtxhgmogsFpFbPW1MHRJnjNkJ9gcHxHrYnrribhFZ7jTffbKpKiJtgJ7AfBrI56yiWbcMNMacBgzDNmnO8LRBitt4E2gP9AB2Ai961pzaR0QigM+BPxtjDnnanoEzuDcAAAR2SURBVLpCRbMOMcZkOK9ZwJdAX89aVGdkikg8gPOa5WF73I4xJtMYU2iMKQLewcc+axEJxArmJ8aYL5zDDeJzVtGsI0QkXEQii98DQ4CV5V/lM7iub38d8D8P2lInFIuHw2X40GctIoJdfnuNMeYll1MN4nNWj6A6QkTaYWuXYJdO/tQY87QHTXILIjIOOAsbJiwTeAz4CpgItAK2AiONMT4zcFLGPZ+FbZobYDNwW3F/n7cjIoOAn4EVQJFz+G/Yfk2f/ZyLUdFUFEWpAto8VxRFqQIqmoqiKFVARVNRFKUKqGgqiqJUARVNRVGUKqCiqZSLiOSU2L9eRF73lD1l4USQalb83oN2XC8iCZ4qX3E/KpqKRxCLT33/RMQfuB6okmiKSIBbDFLcgk99aZW6RURiRORzEVnobAOd44+LyH0u6VaKSBtnWyMibwBLgJYiMkRE5onIEhH5zPFnRkSeFZHVTsCLf5VSdlMRmSYiv4nIW4C4nN7tpBERecEpf4WIXOEcP0tEvnbJ63WnhjhMRCa6HD9LRKY478uyc7OIPCoic4BRQG/gEyeGZmiJGnBvEZnl8ozeFpFpwIci4u/YutC559tq/AEpbkFFU6mIUJdAukuBJ13OvYqNGdkH+APwbiXy6wR8aIzpCeQCfwfOdQKZLAL+IiJNsK6HqcaYbsBTpeTzGDDHyWcy1gsFAMcegN9jvXK6A+cCL5RwbyzJdKC/4+YKcAUwwRG9U+x0uS7PGDPIGPOxc+5qJ4bmkQqeRS9guDHmKuAm4KBjex/gFhFpW8H1igfQZoFSEUeMMT2Kd0TkemxtCqwQpVhXZACiiv3ry2GLMeZX531/IAWY6+QRBMwDDgF5wLsi8g3wdSn5nIEVRYwx34jI/lLSDALGGWMKscEkfsIKUqkReYwxBSLyPXCxiEwCLgQeAM4sw85iJlRwz2Ux2UVYhwDdRGSEsx8NJAGbqpm34iZUNJWa4AcMKFmjEpECTm7FhLi8z3VNCkw3xowqmbGI9AUGA1cCdwPnlFJ+RT7AUsbx8uybANyFDSq80BiT7QSoKNVOh9wyjpcsK6TEuZLPYrQxZmo5eSn1AG2eKzVhGlbQABCR4hrpZuA059hpQFnNzF+BgSLSwUkbJiIdnf7CaGeJiD9jm9glmQ1c7Vw3DCgtyO9s4AqnvzAGWztdAGzB1pCDRSQaK87FzHJsv4UTNchS7SzjnrIB19r2ZmwzHGwXRllMBe5wQq7hPIfwctIrHkJFU6kJ9wC9nYGL1cDtzvHPgSZOH+gdwPrSLjbG7MaONo8TkeVYceqMFZ2vnWM/Af9XyuVPAGeIyBJs03ZrKWm+BJYDy4AfgAeMMbuMMdv4//buGIVhGAgC4Op5fl66uHPjl7hPIIV/Y3AhG1wkgetczDSqBFct0glOfRrPJ8mc5H2paUtvBwzH+q/Ob6Ykz/Mh6Kjz0Vpbkmw/9iS9H7wmebX+QdsYN8FbMuUIoMBJE6BAaAIUCE2AAqEJUCA0AQqEJkCB0AQo2AHFX9LbW70VrQAAAABJRU5ErkJggg==\n", "text/plain": [""]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["ax = res.plot(x='end', y=['score_tree', 'score_lin'], figsize=(5,4))\n", "ax.set_title(\"Evolution du R2 en fonction\\ndu nombre d'heures d'ouvertures\")\n", "ax.set_xlabel(\"Heures d'ouverture\")\n", "ax.set_ylabel(\"R2\");"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On voit que le $R^2$ pour l'arbre de d\u00e9cision et le mod\u00e8le tendent vers 0 lorsque le nombre d'heures d'ouverture tend vers 24, cela signifie qu'il n'y a plus d'effets de seuil : la dur\u00e9e entre la r\u00e9ception et la commande suit une loi gamma et est compl\u00e8tement al\u00e9atoire. C'est attendu. A l'oppos\u00e9, lorsque le nombre d'heures d'ouverture est tr\u00e8s faible, la dur\u00e9e entre la commande et la r\u00e9ception est quasiment un mulitple de 24h puisque le magasin n'est quasiment jamais ouvert : l'arbre de d\u00e9cision est nettement meilleur dans ce cas. Le pic de $R^2$ est obtenu vers 4-5h d'ouvertures, cela correspond au moment o\u00f9 il y a au pire une seule nuit entre la r\u00e9ception et la commande.\n", "\n", "Pour ce probl\u00e8me, le coefficient $R^2$ ne para\u00eet pas \u00eatre la bonne solution puisque de toute fa\u00e7on, le mod\u00e8le pr\u00e9dit simplement si le paquet sera r\u00e9ceptionn\u00e9 le jour m\u00eame ou le suivant. Ce serait plut\u00f4t un probl\u00e8me de classification binaire."]}, {"cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [], "source": []}], "metadata": {"kernelspec": {"display_name": "Python 3", "language": "python", "name": "python3"}}, "nbformat": 4, "nbformat_minor": 2}