{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Chapter 17: Experiments with Blocks"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import statsmodels.api as sm\n",
"import statsmodels.formula.api as smf\n",
"import seaborn as sns\n",
"from scipy import stats"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Randomised block design"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" yield | \n",
" block | \n",
" variety | \n",
"
\n",
" \n",
" \n",
" \n",
" 1 | \n",
" 296 | \n",
" I | \n",
" 1 | \n",
"
\n",
" \n",
" 2 | \n",
" 357 | \n",
" II | \n",
" 1 | \n",
"
\n",
" \n",
" 3 | \n",
" 340 | \n",
" III | \n",
" 1 | \n",
"
\n",
" \n",
" 4 | \n",
" 331 | \n",
" IV | \n",
" 1 | \n",
"
\n",
" \n",
" 5 | \n",
" 348 | \n",
" V | \n",
" 1 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" yield block variety\n",
"1 296 I 1\n",
"2 357 II 1\n",
"3 340 III 1\n",
"4 331 IV 1\n",
"5 348 V 1"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"oatvar = pd.read_csv(\"data/oatvar.csv\", index_col=0, dtype={'yield':'int','block':'category','variety':'category'})\n",
"oatvar.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Display the data as a table:"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" block | \n",
" I | \n",
" II | \n",
" III | \n",
" IV | \n",
" V | \n",
"
\n",
" \n",
" variety | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
"
\n",
" \n",
" \n",
" \n",
" 1 | \n",
" 296 | \n",
" 357 | \n",
" 340 | \n",
" 331 | \n",
" 348 | \n",
"
\n",
" \n",
" 2 | \n",
" 402 | \n",
" 390 | \n",
" 431 | \n",
" 340 | \n",
" 320 | \n",
"
\n",
" \n",
" 3 | \n",
" 437 | \n",
" 334 | \n",
" 426 | \n",
" 320 | \n",
" 296 | \n",
"
\n",
" \n",
" 4 | \n",
" 303 | \n",
" 319 | \n",
" 310 | \n",
" 260 | \n",
" 242 | \n",
"
\n",
" \n",
" 5 | \n",
" 469 | \n",
" 405 | \n",
" 442 | \n",
" 487 | \n",
" 394 | \n",
"
\n",
" \n",
" 6 | \n",
" 345 | \n",
" 342 | \n",
" 358 | \n",
" 300 | \n",
" 308 | \n",
"
\n",
" \n",
" 7 | \n",
" 324 | \n",
" 339 | \n",
" 357 | \n",
" 352 | \n",
" 220 | \n",
"
\n",
" \n",
" 8 | \n",
" 488 | \n",
" 374 | \n",
" 401 | \n",
" 338 | \n",
" 320 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
"block I II III IV V\n",
"variety \n",
"1 296 357 340 331 348\n",
"2 402 390 431 340 320\n",
"3 437 334 426 320 296\n",
"4 303 319 310 260 242\n",
"5 469 405 442 487 394\n",
"6 345 342 358 300 308\n",
"7 324 339 357 352 220\n",
"8 488 374 401 338 320"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"oatvar.pivot(index = 'variety', columns='block', values='yield')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Make plots of the data:"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.boxplot(x=\"variety\", y=\"yield\", data=oatvar)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEKCAYAAAAIO8L1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAFsFJREFUeJzt3X2QXfV93/H3d2GNRIgjo5VtrAuVxyvHdTsYkw12x50Ug6WwDhNIx07xNOmOhwlOYiNnUj+RcTNJx54pnaR4VklJZeNkXTvIhMY1pawLtVEyeGIcCcQikFstRphrhKUViKBqkQX77R/3rLUsR7t3V3vu2Yf3a+bO3vM7T9+9iP3c33n4nchMJEmarqvuAiRJi5MBIUkqZUBIkkoZEJKkUgaEJKmUASFJKmVASJJKVRoQEbE/Ih6OiN0RsbNoOzci7omIfcXP1xTtERGDETEaESMRcXGVtUmSZtaJHsS7M/OizOwrpj8FfDMzNwLfLKYB+oGNxes64OYO1CZJOoUza9jnVcClxfshYAfwyaL9S9m6tfs7EbEmIs7LzAOn2lBPT09u2LCh2molaZnZtWvXWGaum225qgMigbsjIoH/kpnbgNdN/tHPzAMR8dpi2fXAk1PWbRZtLwuIiLiOVg+DCy64gJ07d1b8K0jS8hIRT7SzXNUB8a7MfKoIgXsi4nszLBslba8YKKoImW0AfX19DiQlSRWp9BxEZj5V/DwIfA24BPhRRJwHUPw8WCzeBM6fsnoDeKrK+iRJp1ZZQETET0XET0++BzYDe4A7gIFisQHg68X7O4B/U1zN9E7guZnOP0iSqlXlIabXAV+LiMn9/GVmfiMi/h64LSKuBX4AvL9Y/i7gvcAocAz4YIW1SZJmUVlAZOb3gbeVtB8GLi9pT+DDVdUjSZob76SWVKmxsTGuv/56Dh8+XHcpmiMDQlKlhoaGGBkZYWhoqO5SNEd13Ci3KA0ODjI6Ojrn9ZrNJgCNRmNO6/X29rJly5Y5709aSsbGxhgeHiYzGR4eZmBggLVr19ZdltpkD+I0jY+PMz4+XncZ0qI0NDTE5HPvJyYm7EUsMTH5H28p6uvry7rvpJ7sBQwODtZah7QYXXHFFRw7duwn02effTbf+MY3aqxIABGxa8r4eKdkD0JSZTZt2kR3dzcA3d3dbN68ueaKNBcGhKTKDAwMUNwLRVdXFwMDA7OsocXEgJBUmZ6eHvr7+4kI+vv7PUG9xHgVk6RKDQwMsH//fnsPS5ABIalSPT09bN26te4yNA8eYpIklTIgJEmlDAhJUikDQpJUyoCQJJUyICRJpQwISVIpA0KSVMqAkCSVMiAkSaUMCElSKQNCklTKgJAklTIgJEmlDAhJUikDQpJUyoCQJJUyICRJpSoPiIg4IyIejIg7i+m/iIjHI2J38bqoaI+IGIyI0YgYiYiLq65NknRqnXgm9UeBvcCrp7R9PDNvn7ZcP7CxeL0DuLn4KUmqQaU9iIhoAL8EfKGNxa8CvpQt3wHWRMR5VdYnSTq1qg8xfQ74BDAxrf2zxWGkmyLirKJtPfDklGWaRdvLRMR1EbEzInYeOnSokqIlSRUGRERcCRzMzF3TZt0AvAX4eeBc4JOTq5RsJl/RkLktM/sys2/dunULWbIkaYoqexDvAn45IvYD24HLIuLLmXmgOIx0HPhz4JJi+SZw/pT1G8BTFdYnSZpBZSepM/MGWr0FIuJS4GOZ+WsRcV5mHoiIAK4G9hSr3AF8JCK20zo5/VxmHpjPvgcHBxkdHT3t36Ed+/btA2DLli2V76u3t7cj+5Ek6MxVTNN9JSLW0TqktBv4zaL9LuC9wChwDPjgfHcwOjrKgw8/ysTZ555urbOKH7eOgu167OlK99N17JlKty9J03UkIDJzB7CjeH/ZKZZJ4MMLtc+Js8/lhbdeuVCbq92qR++su4QVab690WazCUCj0ZjTevYStZjU0YOQlr3x8fG6S5BOmwEhzWC+3+Yn1xscHFzIcqSOciwmSVIpA0KSVMqAkCSVMiAkSaUMCElSKQNCklTKgJAklTIgJEmlvFFOK4aDOEpzY0BoxRgdHeXBRx6ENR3YWfGIrAd/+GC1+zlS7ea1shkQWlnWwMSl0x9wuHR17fAosarjvy5JUikDQpJUyoCQJJUyICRJpQwISVIpA0KSVMqAkCSVMiAkSaUMCElSKQNCklTKgJAklTIgJEmlDAhJUilHc9UrzPe5Cc1mE4BGozGn9XyegbQ4Vd6DiIgzIuLBiLizmH5jRNwfEfsi4qsR8aqi/axierSYv6Hq2rSwxsfHGR8fr7sMSQukEz2IjwJ7gVcX0zcCN2Xm9oj4M+Ba4Obi57OZ2RsR1xTL/asO1Kdp5vttfnK9wcHBhSxHUk0q7UFERAP4JeALxXQAlwG3F4sMAVcX768qpinmX14sL0mqQdWHmD4HfIKfPICRtcCRzHyxmG4C64v364EnAYr5zxXLS5JqUFlARMSVwMHM3DW1uWTRbGPe1O1eFxE7I2LnoUOHFqBSSVKZKnsQ7wJ+OSL2A9tpHVr6HLAmIibPfTSAp4r3TeB8gGL+zwDPTN9oZm7LzL7M7Fu3bl2F5UvSwhobG+P666/n8OHDdZfSlsoCIjNvyMxGZm4ArgG+lZn/GrgXeF+x2ADw9eL9HcU0xfxvZeYrehCStFQNDQ0xMjLC0NDQ7AsvAnXcKPdJ4HcjYpTWOYZbivZbgLVF++8Cn6qhNkmqxNjYGMPDw2Qmw8PDS6IX0ZGAyMwdmXll8f77mXlJZvZm5vsz83jR/kIx3VvM/34napOkThgaGmLyoMjExMSS6EU41IYkdcA999zDiRMnADhx4gR33313zRXNzoCQpA7YtGkT3d3dAHR3d7N58+aaK5qdASFJHTAwMMDkvb9dXV0MDAzMskb9DAhJ6oCenh76+/uJCPr7+1m7dvHfB+xorpLUIQMDA+zfv39J9B7AgJCkjunp6WHr1q11l9E2DzFJkkoZEJKkUgaEJKmU5yC0YjSbTXgOunYso+9FR6CZzbqr0DJlQCxz832+9Hzs27cPmP8T6ebC51hL1TMglrnR0VH+754HuOCclyrf16tOtL6Zv7D/7yvdzw+OnjGv9RqNBofiEBOXTsy+8BLRtaOLxvpG3WVomTIgVoALznmJT/cdrbuMBfOZnefUXYK0Iiyjg7GSpIW0LHsQzWaTrmPPserRO+suZcF0HTtMs/ni7AtK0gKxByFJKrUsexCNRoMfHT+TF956Zd2lLJhVj95Jo/H6usuQtILYg5AklTIgJEmlTnmIKSIunmnFzHxg4cuRJC0WM52D+OPi5yqgD3gICOBC4H7gn1dbmiSpTqc8xJSZ787MdwNPABdnZl9m/hzwdqAzYzdIkmrTzjmIt2Tmw5MTmbkHuKi6kiRJi0E7l7nujYgvAF8GEvg1YG+lVUmSatdOQHwQ+C3go8X03wI3V1aRJGlRmDUgMvMF4KbiJUlaIWa6zPVhWoeUSmXmhZVUJElaFGbqQSyfcSokSXM202WuT0y+iqaNxfuDwDOzbTgiVkXEdyPioYh4JCL+sGj/i4h4PCJ2F6+LivaIiMGIGI2Ikdlu1JMkVWvWcxAR8RvAdcC5wJuABvBnwOWzrHocuCwzj0ZEN3BfRAwX8z6embdPW74f2Fi83kHrRPg72v1FpLYc6dAzqSefz1T1s42OAOsr3odWrHauYvowcAmtu6fJzH0R8drZVsrM5OT/Jt3F65TnNICrgC8V630nItZExHmZeaCNGqVZ9fb2dmxfk8/n3rh+Y7U7Wt+532u+zzdvNptAa5TlufC54/VrJyCOZ+aPIwKAiDiTmf/Q/0REnAHsAnqBP83M+yPit4DPRsTvA98EPpWZx2l9D3pyyurNos2AOA3NZpP/9/wZy+oxnU88fwY/VfzRmYtO/rGZ3Nfg4GDH9rlYjY+P112C5qmdgPibiPg9YHVEbAJ+G/gf7Ww8M18CLoqINcDXIuKfAjcATwOvArYBnwT+Pa1xnl6xiekNEXEdrUNeXHDBBe2UIWkBzDdgDculq52A+BRwLfAw8CHgLuALc9lJZh6JiB3AFZn5R0Xz8Yj4c+BjxXQTOH/Kag3gqZJtbaMVLPT19bXVk1nJGo0GL7x4gE/3HZ194SXiMzvPYdUcD1dIC2mlHG5r50a5CeDzxattEbEOOFGEw2rgPcCNk+cVonXM6mpgT7HKHcBHImI7rZPTz3n+QdJystQOt810o9xtmfmrp7phro0b5c4DhorzEF3AbZl5Z0R8qwiPAHYDv1ksfxfwXlojxR6jNcSHJC06K+Vw20w9iMmxl7YB99HGvQ9TZeYIraHBp7dfdorlk9YVU5IqNt9DJPMxeUVXJw6ReOXTwjplQEw5vPM64K+AB4AvAv+r+GMuaYkaHR3le7t38/oO7GvyrpMju3dXup+nK936ytTOOYhPR8S/AzbTOuzzJxFxG3BLZj5WdYGSqvF64NrSiweXplvau/pec9DWLaVFj+Hp4vUi8Brg9oj4jxXWJkmqUTtDbWwBBoAxWpe3fjwzT0REF7AP+ES1Jep0/eBoZ26U+9Gx1veN1509Uel+fnD0DN5c6R4kQXv3QfQA/3LKoH1A6/LXiHDE10Wuk8NL/Lg4GblqQ7XDS7yZzv5e0krVzjmI359hno8eXeQcXkJlms0mz7O8jtsfAI7OYwgWnVoHhrWUJC1F7RxikrTMNBoNjoyNLburmNY4BMuCsgchSSplQEiSShkQkqRSnoOQtKI5LtWpGRCSVrTR0VEeeXgva86e9UnKp23ix62LAn742OFK93Pk2MEF2Y4BIWnFW3P2a3n3W66pu4wFc+/3ti/IdjwHIUkqZUBIkkoZEJKkUsv2HETXsWdY9eidle8nXvgHAHLVqyvdT9exZ6Ajj3eZ/1Ud871Cw6eASYvTsgyITo70uW/f8wBsfFPVf7xfv+hHMF29enXdJUhaQMsyIBzB9PT4bV4SLNOAkDS7p+nMcN+TV/yvrXg/TwNrKt7HSmNASCtQJw9XHirOTa3ZWO2DpNYwv9+r2Wzy3LHnF+zegcXgyLGDZHP8tLdjQEgrkIdh1Q4DQtKK1mg0iOOHl92d1Osbp39Qz4CQZuAlv1rJDAipAl7yq+XAgJBm4Ld5rWQOtSFJKlVZQETEqoj4bkQ8FBGPRMQfFu1vjIj7I2JfRHw1Il5VtJ9VTI8W8zdUVZskaXZV9iCOA5dl5tuAi4ArIuKdwI3ATZm5EXgWuLZY/lrg2czsBW4qlpMk1aSygMiWo8Vkd/FK4DLg9qJ9CLi6eH9VMU0x//KIiKrqkyTNrNKT1BFxBrAL6AX+FHgMOJKZLxaLNIH1xfv1wJMAmfliRDxH6+78sSprlNQeL/ldeSo9SZ2ZL2XmRUADuAT4x2WLFT/LeguvGCgmIq6LiJ0RsfPQoUMLV6ykSqxevdrLfpeojlzmmplHImIH8E5gTUScWfQiGsBTxWJN4HygGRFnAj8DPFOyrW3ANoC+vr7qRxqTBHjJ70pUWUBExDrgRBEOq4H30DrxfC/wPmA7MAB8vVjljmL674r538rMjgWA3WdJerkqexDnAUPFeYgu4LbMvDMiHgW2R8RngAeBW4rlbwH+a0SM0uo5LImBUew6S1quKguIzBwB3l7S/n1a5yOmt78AvL+qembjt3lJejnvpJYklTIgJEmlDAhJUikDQpJUyoCQJJUyICRJpQwISVIpnygnacU7cuwg935ve+X7OfrCswCcs+o1le7nyLGDrGftaW/HgJC0ovX29nZsX/v2tYaXW/+m0//jPZP1rF2Q38uAkLSidXIUhcl9DQ4Odmyfp8NzEJKkUgaEJKmUASFJKmVASJJKGRCSpFIGhCSplAEhSSplQEiSShkQkqRSBoQkqZQBIalSY2NjXH/99Rw+fLjuUjRHBoSkSg0NDTEyMsLQ0FDdpWiODAhJlRkbG2N4eJjMZHh42F7EEmNASKrM0NAQmQnAxMSEvYglxoCQVJl77rmHEydOAHDixAnuvvvumivSXBgQkiqzadMmuru7Aeju7mbz5s01V6S5MCAkVWZgYICIAKCrq4uBgYGaK9JcGBCSKtPT00N/fz8RQX9/P2vXVvuoTS2sygIiIs6PiHsjYm9EPBIRHy3a/yAifhgRu4vXe6esc0NEjEbE/4mIX6yqNkmdMzAwwIUXXmjvYQmq8pnULwL/NjMfiIifBnZFxD3FvJsy84+mLhwRbwWuAf4J8Abgf0fEmzPzpQprlFSxnp4etm7dWncZmofKehCZeSAzHyjePw/sBdbPsMpVwPbMPJ6ZjwOjwCVV1SdJmllHzkFExAbg7cD9RdNHImIkIr4YEa8p2tYDT05ZrcnMgSJJqlDlARER5wD/DfidzPwH4GbgTcBFwAHgjycXLVk9S7Z3XUTsjIidhw4dqqhqSVKlARER3bTC4SuZ+dcAmfmjzHwpMyeAz3PyMFITOH/K6g3gqenbzMxtmdmXmX3r1q2rsnxJWtGqvIopgFuAvZn5n6a0nzdlsV8B9hTv7wCuiYizIuKNwEbgu1XVJ0maWZVXMb0L+HXg4YjYXbT9HvCBiLiI1uGj/cCHADLzkYi4DXiU1hVQH/YKJkmqT2UBkZn3UX5e4a4Z1vks8NmqapIktc87qSVJpQwISVIpA0KSVMqAkCSVMiAkSaUMCElSKQNCklTKgJAklTIgJEmlDAhJUikDQpJUyoCQJJUyICRJpQwISVIpA0KSVMqAkCSVqvKJcpK0LA0ODjI6Ojrn9fbt2wfAli1b5rReb2/vnNdZCAaEJHXI6tWr6y5hTgwISZqjOr7N18FzEJKkUgaEJKmUASFJKmVASJJKGRCSpFIGhCSplAEhSSplQEiSSkVm1l3DvEXEIeCJuusAeoCxuotYJPwsTvKzOMnP4qTF8Fn8o8xcN9tCSzogFouI2JmZfXXXsRj4WZzkZ3GSn8VJS+mz8BCTJKmUASFJKmVALIxtdRewiPhZnORncZKfxUlL5rPwHIQkqZQ9CElSKQPiNEXE0bprWAwmP4eI2BARe+qupw5ln0FEXBoRd9ZbWedN+Swej4ifnTbvcxHxiXoq67yI2BERvzit7Xci4j/XVVO7DAhJVdoOXDM5ERFdwPuAr9ZWUefdypTPoHBN0b6oGRCSqjT9j+MvAPszczHc4NoptwNXRsRZ0OphAm8A7quxprYYEJIqk5kjwEREvK1oWhLfnBdSZh4GvgtcUTRdA3w1l8AVQgaEpKrdClwTEWcCVwF/VXM9dZjak1oyIWlASKrarcCvAu8BRjLzYM311OG/A5dHxMXA6sx8oO6C2mFASKpUZj4GHAb+A0vkm/NCy8yjwA7giyyhz8CAkNQJtwJvAb5WdyE1uhV4G60ru5YE76SWJJWyByFJKmVASJJKGRCSpFIGhCSplAEhSSplQEhtONUotcVInXN+vnBE/EFEfGxhqpOqYUBIkkoZEFL7zoyIoYgYiYjbI+LsqTMj4gMR8XBE7ImIG6e0XxERD0TEQxHxzekbjYjfiIjhiFjdiV9CateZdRcgLSE/C1ybmd+OiC8Cvz05IyLeANwI/BzwLHB3RFwNfBv4PPALmfl4RJw7dYMR8RFgM3B1Zh7v0O8htcWAkNr3ZGZ+u3j/ZWDLlHk/D+zIzEMAEfEVWs8+eAn428x8HCAzn5myzq8DTVrhcKLq4qW58hCT1L7p49JMnY5TrBMl603aA2wAGqdXllQNA0Jq3wUR8c+K9x/g5U8Eux/4FxHRExFnFPP/Bvi7ov2NANMOMT0IfAi4ozhEJS0qBoTUvr3AQESMAOcCN0/OyMwDwA3AvcBDwAOZ+fXikNN1wF9HxENMexZzZt4HfAz4nxHR05lfQ2qPo7lKkkrZg5AklTIgJEmlDAhJUikDQpJUyoCQJJUyICRJpQwISVIpA0KSVOr/AyCirfnXHbODAAAAAElFTkSuQmCC\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.boxplot(x=\"block\", y=\"yield\", data=oatvar)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.catplot(x='variety',y='yield',hue='block', data=oatvar, kind='point')\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAFgCAYAAABKY1XKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzsnXdcVFf6/99nhl5FaYooNsQKKqjY0RRjNzFqjJqySTbNX8pms9nyTczuJms2m2Td9MRsktUYS4yxG41dUQQVRUWUKCoqIkrvzNzfH2cYRenMMKDn/XrxYu6959z7zDDc555znufzCE3TUCgUCoWiOnS2NkChUCgUTR/lLBQKhUJRI8pZKBQKhaJGlLNQKBQKRY0oZ6FQKBSKGlHOQqFQKBQ1opyFQqFQKGpEOQuFQqFQ1IhyFgqFQqGoETtbG9AQRo8erW3cuNHWZigUCkVNCFsb0FCa9cgiIyPD1iYoFArFHYFVnYUQIkUIkSCEiBdCxJn2tRRCbBZCnDL99jLtF0KI/wghkoUQR4QQfa1pm0KhUChqT2OMLKI0TQvTNC3ctP0asEXTtC7AFtM2wH1AF9PPU8CnjWCbQqFQKGqBLaahJgLfml5/C0y6Yf//NMk+oIUQorUN7FMoFArFTVjbWWjAJiHEASHEU6Z9fpqmXQIw/fY17Q8Azt/QN9W0rwJCiKeEEHFCiLgrV65Y0XSFQqFQlGPtaKjBmqZdFEL4ApuFECeqaVtZtMAtxTY0TfsC+AIgPDxcFeNQKBSKRsCqIwtN0y6afqcDK4H+wOXy6SXT73RT81Qg8IbubYGL1rRPoVAoFLXDas5CCOEqhHAvfw3cAxwFVgOPmJo9AqwyvV4NzDZFRQ0EssunqxQKhUJhW6w5DeUHrBRClF9nsaZpG4UQscAyIcRvgHPAg6b264ExQDJQADxmRdsUCoVCUQes5iw0TTsNhFay/yowqpL9GvCctexRKBQKRf1p1hncCoVCoWgcmrU2VGOTlZXFqVOnMBgMtGvXjjZt2tjaJIVCoWgUlLOoBQaDgY0bNxIXF4ecLZN06NCBBx98EBcXFxtap1DUHaPByLlj18hKL8DV05GgUG/sHfS2NkvRhFHOohZs27aN2NjYW/afOXOGpUuX8uijj2JayFcomjxpp7P5+cuj5GUWm/c5Otsx/OGudAn3s6FliqaMWrOogaKiImJiYqo8fvbsWc6fP1/lcYWiKZF7rYg1/4mv4CgAigvL2PzVMS4mZ9nIMkVTRzmLGrh06RKlpaXVtklJSWkcYxSKBnJ0RyolRYZKj2kaHPr5bCNbpGguKGdRA7WZXtLp1MeoaB5cOFn9yCG1huOKOxd1l6uBgIAAnJ2dq23TuXPnRrJGobAuOrX0pqgC5SxqwN7enqFDh1Z5vGvXrvj7+zeiRQpF/SgpLKO0uPIpqHLa92zVSNYomhsqGqoWREZGomkaO3fupLi44sKgn5+KHlE0fTJSc9n4xVGy0wurbRd2T7tGskjR3FDOohYIIRg8eDDh4eGcO3eO/Px81q1bR2lpKTExMQwcOFDlWiiaJJqmkRh9iZ1LTmIoNQLg7G6Po6s9WWkFt7RPPZGJbzuPxjZT0QxQzqIOODo60qVLFwCuXbtmHmns2bOHu+++28bWKRQVKS02sOP7JJL2pZn3BXZvyd2PdcfZ3YEr53PJTi9EMxrZ8u0JDGVGYtecoVMfHzx91MOPoiJqzaKeREZG4uTkBEBMTAy5ubk2tkihuM61S/ksnxdndhRCQP/xHRj3fCjO7g4A+AS607mfL10i/IkYFwRAWamR7d8lVVAqUChAOYt64+zszODBgwEoKytj165dNrZIoZAkxaSx/B+xZF7KB8DZw4EJL4QRMbYDuirCncLubkertm6AnIo6sTet0naKOxflLBrAgAEDcHV1BSAuLo6sLBWjrrAdZSUGti06wS9fH6esRK5PtOnSgml/jqBtSMtq++r1OqJmhlCeVrTnh1MU5JRY22RFM0I5iwbg4OBgDqs1Go3s2LHDxhYp7lSyLhfwwz8PcHz39UrE/e5rz8QXw3D1dKzVOfyCPOg9UlY2Li4oY/eyk1axVdE8Uc6igYSHh+PhIaNH4uPjycjIsLFFijuN5APpLPtHLFdT8wBwcrVn3POhDJzYCZ2+bv/iAyZ0xL2VXIs7FZdOSoL6Piskylk0EDs7O0aMGAHIMMVt27bZ1iDFHYOh1MjOpSf5+cujlJr0nvw7ejD1zxH1Tq6zd9QzYkZX8/aOxUmUFJVZxF5F80Y5CwsQGhpKy5ZyTvjYsWNcunTJxhYpbndyMgr58V8HSNiWat4Xdlcgk37XF/eWTg06d7serQgeIJNN8zKLiVl1ukHnU9weKGdhAfR6PVFRUeZtNbpQWJMzh6+w7O1Y0s/KcG1HFzvue7oXg6d0QV/HaaeqGDKlC06u9gAc2Z5K2ulsi5xX0XxRzqIOlBYbOHvsKmeOZNwSKdKjRw+z9MfJkydVjQuFxTEYjESvSGb9pwkUF8ipId/27kz9UwQdw3wsei1ndweGTJUJqGiwbZFM2lPcuagM7lqgaRrxm88TtyGFkkL5T6rTC0IiWzN0ahfsHPTodDpGjhzJ999/D8CWLVt49NFHbWi14nYiL7OIn788VuEJv9eItgx+oDN6e+s88wX39+NkTBrnjl/j2sV8Dm06S/iYDla5lqLpo0YWteDI1lSif0w2OwoAo0Hj+O6LbPk20bwvODiYgIAAQBZEOn1azfUqGs65Y1dZ+las2VHYO+m554keDJsebDVHAVITbfiMrtg5yGvErk8hMy3fatdTNG2Us6iBshIDsevPVHk8+UA6Vy/IkEUhBKNGjTIf27Jli5JNUNQbo1EjZvVp1nx0mKI8Wa2xVYAbU/8Y0Wi1sj28nRkwoaO0p0xj26ITaEb1nb4TUc6iBtLP5lCcX33o4NmjV82vO3bsSIcOcqh+4cIFkpKSrGqf4vYkP7uY1fMPEbc+BUz35u6DWzPlD/1o4de4In+9Rwbi294dgEvJ2Ry7IfFPceegnEUNGA01P0Xd3GbkyJHm11u3bsVoVAuDitqTmpTJ0rdiuZAk5WPsHHSMerQbUbO6Yeegb3R7dDpB1KwQhElXau+PyeRnFdfQS3G7oZxFDfi09zDP2VZFQHCLCtuBgYEEBwcDkJ6ezrFjx6xmn+L2QTNqxK1PYfW/D1Foirbz8ndhymvhhAxsbVPbvNu60+duWRippMjAziVKCuROQzmLGnB0tqPXiLZVNxAypPFmbhxdbNu2DYOh+nKWijubwrwS1n50mJjVpylf5goe4MeU18Jp1cbNtsaZiBgbhKePrEd/Ov4Kvx5Kt7FFisZEOYtaMHBiR3oMC4DK1J01WP9pAlfOVaxn4e/vT8+ePQFZKCk+Pr4RLFU0Ry4lZ7H077GcO34NAL2dVIC969HuODg1neh2Owc9I2aGmLd3LjlJcUGpDS1SNCaiOUfrhIeHa3FxcY12vZyMQs4evYqhzEjrzp4c3XmRE9FS2sPZ3Z7Jv+uLl7+ruX1GRgYff/wxmqbh4eHBnDlzsLe3bzR7FU0bTdOI/+U8e1f+ao4w8vRx5t6neuIT6G5j66pm68JEEvfI732PoW0Y8XBIDT0UVP6o2axQI4s64OHtTK8RbQm7qx1+QZ5EPdyVDqHeABTmlrJ6fjy514rM7b29vQkLCwMgJyeHAwcO2MRuRdOjKL+UDZ8lEL0i2ewoOvX1ZeqfIpq0owAYdH9nnD1ktb1juy5y8VSmjS1SNAbKWTQAnV7HPU/0IKCrFyBF11bPj6cw97oUyPDhw9HrZQTLrl27KC5WUSR3Oulnc1j2dixnDkv5b51eMHRaMPc+2QMH56Yz7VQVTq72DJsWbN7etiiJslK1Jne7o5xFA7Gz1zPmmV7mOPSsywWs+fCwOdu7RYsW9OvXD4D8/HxiYmJsZqvCtmiaRsL2VFa8e4Dcq3IE6t7Sift/34/eUW0RovnMVHTq60NQbzmqzrpcwIENZ21skcLaKGdhARyc7Bg/Jwwvf5ksdeVcLus+OUJZiXzaGjp0qHmtIjo6msLCQpvZqrANJYVlbFpwjJ1LTmIsk9NOQb29mfrnCPyCPGxsXd0RQjD8oWDsneSo+eDGs2YlA8XtiXIWFsLJzZ4JL4SZawlcPJXFzwuOYTQYcXd3Z8CAAQAUFRURHR1tS1MVjUxGai7L/hFL8gEZaip0gkEPdGbMM73MMuDNETcvJyIndQKkNMm2RScwKimQ2xblLCyIm5cTE14Iw9ld3gBSjmSwdaHU0hk0aBCOjrIW8r59+8jLU09htzuaJsUmf3jnANnpcjTp2sKRyS/3oc/d7ZrVtFNV9BwWgH9HTwAun8khYXtqDT0UzRXlLCxMCz8Xxv+/MBxMw/OkfWnsXn4KZ2dnBg0aBEBpaSm7d++2pZkKK1NabGDLN4myDkSpTNps170l0/4cQevOLWro3XwQOkHUzBB0eun49q06XSEiUHH7oJyFFfAJdGfs86HYmeSjj2xLJW59CgMHDsTFRa5rxMbGkp2tqo/djly7mM/yeXEkxaQBIAQMmNCRcc+H4uzuYGPrLE/LNq70G90egLJiAzsWJym15dsQ5SysRJvOLbj3qZ7oTOJr+9ecISn6CkOHDgXAYDCwY8cOW5qosAJJMWksnxdL5iVZ98HZw4EJL/YhfEyQWYjvdqTf6CBzgMfZo1dJjlNSILcbyllYkaBe3ox6rJs5d3PX0pN4aoG4u8sw20OHDnH16tVqzqBoLpSVGNi26AS/fH2cshI57RQQ3IJpf46grSkP53ZGby8lSsrZteykuQaH4vZAOQsrExzhz/Dp1xOYti86Rc/OMu9C0zS2b99uI8sUliLrcgE//PMAx8vrPAgIHxPEhBf74OrpaFvjGpHWnVvQc7isFFmYW8qeFadsbJHCkljdWQgh9EKIQ0KItabtb4QQZ4QQ8aafMNN+IYT4jxAiWQhxRAjR19q2NRY9h7dlwARZEEkzapz+xYiHm4wgSUhI4PLly7Y0T9EAkg+ks+wfsVxNldFtTm72jH8+lAETOpqnIO8kIid1wrWFdJAn9qZxPvGajS1SWIrGGFm8ACTetO/3mqaFmX7K5VjvA7qYfp4CPm0E2xqNfvcFEToqEABjKegvtzEf27p1q63MUtQTQ6mRnUtO8vOXRyktksmXrTt5Mu3PEbTr0crG1tkOB2c7hj90w0j6uxOUligpkNsBqzoLIURbYCywoBbNJwL/0yT7gBZCCNtWfLEgQggGP9CZkEh/APS53tgbpUJtUlISqakqPr25kJNRyI//OlAhp6DP3e2Y+HIf3LycbGhZ06BDqA+d+voAkJNRROyaqmvYK5oP1h5Z/Bt4Fbi5OtBbpqmmD4QQ5ZO6AcD5G9qkmvZVQAjxlBAiTggRd+XKFasYbS3KY9I7hHojEDhntzcfU6OL5sGZw1dY9nYs6Wdl/RJHFzvGPNOLQQ90Rq9XS4DlDJ0WjKOLFEWM33L+lnoviuaH1b7dQohxQLqmaTfrcv8RCAEigJbAH8q7VHKaW4K1NU37QtO0cE3Twn18fCxpcqNwXam2BQ7FrbArlVXQTp8+zZkz6gmsqWIwGNmzIpn1nyZQXCBFIn3buzP1TxF0CG1+30Nr4+rpyKD7OwNynW7rwkSMlVSUVDQfrPkoNBiYIIRIAZYAI4UQizRNu2SaaioGvgb6m9qnAoE39G8LXLSifTZDKtX2xq+9B665Qeb9W37ZopKZmiB5mUX89N4h4jefM+/rFdWW+1/ph4e3sw0ta9p0G9zaXJ8+43we8VvO19BD0ZSxmrPQNO2Pmqa11TQtCJgObNU0bWb5OoSQwjiTgKOmLquB2aaoqIFAtqZpl6xln61xcLJj3JxQfFsGYF8iI6NSL6RyIjHJxpbduRgMRnKuFlbIDzh77CpL/x5L2mmZbW/vpOfeJ3sybFowens17VQdQghGPByC3k5+TrFrzpB9pcDGVinqiy0qrXwnhPBBTjvFA0+b9q8HxgDJQAHwmA1sa1Sc3RyY+EIfvvvXVS4hy8Ou+XEDwcFd0NvpbWzdnYPRYOTAxrMkbE+lMFc6ioCuLXBv5cyJvZfMk6Gt2rox+smetPBzsaG1zYsWfi5EjAti30+nKSs1sv27JCa8EHZbiCjeaaga3E2ArMsFfPbhVxTZyWzuEJ9Ipj1zz20tD9FU0DSNzf89zqnY6nNdug9tw9AHu2DnoJx4XTEYjCx/O85c72LUI90IibxtAh1rS7P/Z1bj6CZACz8XJtx/n3n71KV4di5XYmyNQdrpnGodhRBw12PdiXo4RDmKeqLX64iaFUL5YGL3D6coyCmpvpOiyaGcRROhe9/OdGjXBQCDXSFx+w4Stz7FtkbdASQfqH5EoWkQ2K1lI1lz++IX5EHvkTJ+pTi/jN3LlRRIc0M5izqSW5JLZlGmVZ76x4y/B2Earea7nSVmza+qmIyVKSmqObu4pKisESy5/ek/voO5kuSp2MukJGTY2CJFXVDOopbEpsXy6MZHGfT9IIYtHcaEnyaw8tRKizoNHx8feof2BsCoL6bIJY2dS06a6yIoLEvutSLSz+ZU28bR1c58g1M0DAcnO4Y/3NW8vWNxknLEzQjlLGrBrtRdPLnpSQ5cvp5fmJKTwuvRr/P5kc8teq0RI0ag08k/S4HrOTQMbPk2UT2FWRCjwcjhLef5/s0Yrl3Ir7Ztz6EB5tBPRcNp36MVwf39AMjLLCZm1WkbW6SoLeq/oAaMmpF5++dh0Cqfrvjs8GekF1iu0IuXlxf9+kkJc6O+hELXi2hGjY1fHOXiqSyLXedOJf1sDj+8c4Ddy09RWiz/pm5ejtg53Pqv0DHMh4hxHRrbxNueIQ92wclV1qk/sj2VtDOqYmRzQDmLGjh+9Tjncs9VedygGdh8drNFrzl06FDs7GQKTLHnBYyiDEOpkXUfH1YaO/WkpLCMXUtP8sO8OPNnqLMTRIzrwMN/Hcgj/xjM4CmdCR7gR8/hAUx6uQ+jf9tTjSqsgLO7A0MelFIgaLBt4QkMZUoKpKlji6S8ZkVOSfVz2rVtUxc8PDzo378/0dHRlBlLcA3OojDJm5IiA2s+jOf+V/qpxLBaomkapw9dYdfSk+RnXw/XDOjaguEPdcXLXyr/2tnrCburna3MvOMIHuBP0v7LnD9+jWsX8zm06SzhY9QorimjHptqoJNnJ3Si+o8puEVwtcfrw5AhQ3BwcAAgrTCZtr1kKdbC3FJWzT9EXmaRxa95u5GTUci6T46w8YujZkfh5GbPqEe7MfHFPmZHYUbToCQfDKocqLURQjBiRlfz9F/s+hQy06pfP1LYFuUsasDP1Y+72t1V5XF/F3+GBQ6z+HVdXFyIjIwEoKSkBMeOVwnoKkXZ8q4Vs3p+PIV5KrGpMgwGIwc3neX7v8ZwNuF6jfNug1rz8NyBhAxsXVFuwmiE/V/Ch/3g7TbwVmtY/ihkqFwAa+Lh7cyACR0BMJZpbFt0As2oElGbKspZ1ILXI1+nR6selR4bHDAYe529Va4bGRmJs7NUNY07EMfQmUH4tpcjjMy0AtZ+eFiFHt5E2ulslr8dx94ff6WsRM6De/m7MPl3fRg5uxtObpX8rTb8Hta/Atd+ldvGUji2EhaMgvSbizwqLEnvqLbm7/Sl5GyO77kthaZvC5SzqAWejp4sHLOQd4e/y9iOYxnRdgSOelmzac2va0jNtU7inJOTE0OGDAGgrKyMvTHRjJsTipe/XK9IP5vL+k+PUFaqylYWF5SyY3ESK949YNYg0tvrGDCxI9P+0p82Xbwq73gxHmKrKORYlA2b/mIlixUg67tEzQox66BFr0gmP6vYxlYpKkM5i1pir7NndNBo5g2dx4ejPmROnzkAlBhLeP/A+1a7bkREBG5uskDSwYMHKSzJY/z/C8OtpXRWF5Ky2LTg2B1bWEbTNE7FXWbx3BiO7rxgVogN7N6Sh17vT/h9QdVHNCUsr/4CyVug4JrlDFbcgndbd/rcLYMLSooM7Fx60sYWKSpDOYt6MiNkBkEeQQBsPruZ2LRYq1zHwcGBYcPkmojRaGT79u24t3Ri4gt9cHaXUypnDmfckfO92VcKWfvhYTYtOGYWpnP2cODu33Rn/JxQPH1qETFWmFlhs4ybyzNqt7RRWJ6IsUF4+sgp19OHrnD6UPMqmXwnoJxFPbHX2/P7iN+bt+ftn4fBaJ3poL59+9KihVzcPnLkCOnp6bTwc2H8nDAcnKQS6om9aez5IfmOUKo1lBk5sDGF7/8aw7nj15/6ewxtw4w3BhAc4V+7egmaBqUFaMBKN1ceaONPnw7tGNC+LX/2bsl5Oz04uIFHG+u9GQUAdg56RswMMW/vWJJEcYGKSmtKKGfRAIa1HcbggMEAnMw8yYpTK6xyHTs7O0aMGGHe3rZtGwA+7dwZ+1xvc8W2w1vPc2BDilVsaCpcTM5i2dux7PvpNIZSOfXWso0r9/++HyMeDjFnBtdIfgYsnQnHVvKvli143acVJx1lqHKhTsdqdzdmtPHnTK9JYK9KpzYGbbt60W2QrHNRkF3C3pW/2tgixY0oZ9FAXg1/FTshcxs/OvSRxRP0yunduzfe3t4AJCYmcvGijBpp08WL0U/1RGdaIIxZfea2VKotyi9l28JEVv7rINcuynh8O3sdkZM7MfXPEbTu5Fn7k51YD58MhBNrSbK353+eHnKUcRNZej3vFZ2x1FtQ1IJBD3TG2UM67WO7LiqJmyaEchYNpGOLjkwPmQ5AZnEmnx3+zCrX0el0REVFmbe3bt1qfh3Uy5uRj3Qzb+9cepKT+28PpVpN00iKSWPx3H0c33O9JHv7nq146I0B9L23PXp9Lb/GRTmw6jlY8hDkyznxtd6mim2VTVtpGjvzz5GddqShb0NRS5xc7Rk6tYt5e9uiEyrar4mgnIUFeDr0aVo4yjWF7xO/53S2dZQ0u3Xrhr+/PwDJycmcPXvWfKzrAH+GTjNlkmuw5Zvmr1SbdbmA1fPj+eXr4+ba2C6eDtz7ZE/GPtcbD+86TA+djYbPBsOhRdf39X2EzG5jqu4jBJoQZG38faUjD4V16NzPl6DechSddbmAAxvO1tBD0RgoZ2EBPB09zaG0ZVoZ/4r9l1Wuo9PpGDlypHl7y5YtFRa0e0e1pf94qa9jbMZKtYZSI7HrzrDkb/tJPWGKRBLQa0RbZswdSOd+vrVbwAYoK4ZN/wdfj4EskyCkqy/MWAYT/kOQV5dquzsbjfiejYED39T/DSnqhBCCYdODsXeUwRsHN541584obIdyFhbigS4PEOwln+x3XdjFztSdVrlOly5dCAyU5SnPnTtHcnJyhePhY4LoPbItQLNUqr1wMpMlf9/P/jVnzEqk3oFuTHk1nGHTg3F0roP2ZVoCfDECov+DOSC22wR4dh8E3wvApM6Tqs3An5hXgLOmSYeTfaGe70pRV9xbOhE5uRMgH3y2LTqB8Q4LDW9qKGdhIfQ6PX+I+IN5+93Ydyk1Wj70TwjBqFGjzNtbt26tMLoQQjBkShe6DpTTVeVKtVmXCyxuiyUpzCthy7fH+en9Q2Zb7Rz1DJ7SmQdfC8evg0ftT2Y0wK734YsoSD8u9zl6wuQvYOr/wLWVuam7g7t5CvFmBIIZne+XGyW5sPYlNR3ViPQcFoB/R/l3v3wmh6M7br/AjeaEchYWpH/r/mbRwZScFJacWGKV6wQFBdGpk3zqunTpEomJFfWLhE4wclaIed63MLeU1fPjm6RSraZpJEZfYvEbMZzYe31RPqi3NzPeGEDYXe3Q1XYBG+Daafj6PtjyptR4AugwDJ6NhtBptyxkf374c64UysXuALcAwv3C6eAhp/I0NL5ycwBv01rQqZ9rzvhWWAyhE0TN7IZOL/9me386Te61pvcdvlNQzsLCvBz+Mg46Gfr3afynXCuyjlTEjWsXW7duxWisKPeh0+u498keBATLp+bca0VNTqk2My2fn94/xNb/JVKUL2/sri0cue/pXox9tnfdal9rGsR9DZ8OgfMxcp+dE4yeB7NWgWfbW7okXUvi66Nfy+vau/LN6G/4evTXLBu/jNauMkpq9el1JEa9CpiczIY/QJ7KLm4sWrZxpd/o9gCUFRvYsTjpjkg8bYooZ2FhAt0Dmd1jNgC5pbl8dOgjq1wnICCAkBCZ8ZqRkUFCQsItbezs9Yx5pjc+7ZqWUm1ZqYGY1adZ8rf95gV4ISB0ZCAz5g6gY5hP3U6YmwaLp8LaF6HUVBOhdRj8dicMfAZ0t37Ny4xlvB79OmWa/Cxe6vsS/q5y6s7JzokX+r4AyNHFv1I3ovV/SnYsvCZVahWNRr/RQWbxzLNHr5IcZ7kyxorao5yFFXii1xP4OMsb3opTK0i6lmSV69yYd7Ft2zbKym51Ag7OdoyfE2qurCeVahNsFrt+PvEaS/66n7j1KRgN8gnRp507D/4xgiFTu+DgVMfijcd+gk8i4dQmuS30MPwP8MQv4NO1ym4Ljy/k+FW5ntHXty8Pdn2wwvH7OtxHz1Y9Adiftp/tIVHQwlRJ79hKSFxTNzsV9UZvryPqBimQXctOUpSnpEAaG+UsrICrvSsv9nsRAKNm5J3Yd6wydPbz86NXr14AZGVlcejQoUrbObs7MOGFMNy8ypVqMxtdqbYgp4TN/z3G6vnxZF8pBMDeSc/QaV2Y8lq4efRTawqzYMWTsPwR+bQP0Koz/GYzRP0J9FVHOJ3LOcfH8R8D4KBz4M1Bb95SDVEndBW0v94//Aml4/59vcG63ymBwUakdecW9BwWAMg1uD0rVGGqxkY5CysxruM4ennLG3lsWiy/nPvFKtcZMWIEOtM0y86dOyktrfyJy72lExNeCGt0pVrNqHFs1wUWz93Hyf2Xzfs79fFhxhsD6R0VaJYqqTW/boNPB0HCsuv7+j8Fv90FbftVb4+mMXfvXIoNsmbCM2HPEOQZVGnbvn59ubv93YAMWFhWmgZ95RQjeZfh5z/XzW5Fgxg4uROuLeQDz4m9aZxPVNLxjYlyFlZCJ3S8GvGqefu9uPfMNyhL0qpVK/r06QNAbm4usbFVS6V7+bsyfk4Y9jcq1a6wnlJDVYITAAAgAElEQVTt1Qt5rHzvINu/S6K4QE6RubV0ZOyzvRn9217mkU6tKSmA9a/CwkmQY8p5cG8Ds1bCmHfBoWZJ8hWnVpjl5ENahvBIj0eqbf9S35ew08mpsc8Of0b2iFfB3SQREv8dJFvnIUBxK47Odgybfr3e/fbvTlBaoqRAGgvlLKxImG8Y4zqOA+BC3gX+d+x/VrnOsGHD0OulA9i1axdFRVWHF/q0c2fsszco1W45b3E5hdISA3tX/sqyt2K59Gs2IMMg+9zdjhlvDDSH9NaJCwfg82Gw//Pr+3o9KENiO42sut8NpBek836cLFSlF3reHPRmjSVxAz0CmREyA4Cs4iy+TFoK4z643mDNi1DcfJIemzsdw3zo1FeuB+ZkFBG7Rgk9NhbKWViZF/u+iLOd1DD6MuFL0gssH8nh6elJREQEAIWFhezbt6/a9gHBXox+sqe5lGXM6tMWS3g6e+wqS/4aw8Gfz5ozbv06eDD1T+EMeqCzWcKh1hhKYds/YMHdcNU0T+3sBVO+hgcWyNe1QNM0/r7v7+SWyhv77B6z6d6qe636PtX7KTwdpart4hOLOd+6J/ScIg9mn4df3qzbe1I0iKHTgnEwZfLHbznfrBQKmjPKWVgZP1c/ftPzNwAUlhUy/+B8q1xnyJAh2NvLp+S9e/dSUFB9xnZQb29G3aBUu2PJSU7G1l+pNj+7mJ8XHGXth4fJyZAjGwdnO4bP6MoDv++Hd9s6LmADXDkJX90NO+aBZppu6Hw3PLMXet5fp1NtOruJbedlHZB27u14NvTZWvf1dPTkmdBnACg1lvLBwQ/gvnfAxZQJHvulFCpUNAquno4MfqAzINfEti5MvGPLCjcmylk0Ao/0eIQ2rrLa2upfV3PkiuUlr93c3IiMjASguLiYPXv21NhHKtWahPQ02PJ13ZVqNaPG0R2pLH5jX4X49y7hvsyYO4CewwLMI5haYzTCvs/g86Fw0RThZe8ip38eXg4eret0uuzibN6Oedu8PXfQXJzs6pDwB0ztOrVCGd1D+efhvn9eb7DqeSgtrNM5FfWn2+DWtOkiE04zzudxeIuSArE2ylk0Ak52Trwc/rJ5+53972DULP8kFBkZiZOTvAnGxMSQm1vz8Lx3VGAFpdqfvzjKxeTaKdVmpOay4t0D7Pj+JCVF8snfw9uJ8XNCueeJnrh61nEBGyA7VS5gb/wDlJnWXtr2h6d3Q/jjldedqIF3Y981Z9JPCZ5ChH9Enc9hr7PnpX4vVTinscdk6DpW7rj2K2x7u4reCksjhCBqZgh6O3kL27/mtDkkW2EdlLNoJO5pfw/9/GRY55GMI6w7vc7i13B2dmbwYFnmtaysjF27dtWqX/iYIHpHSTmMslIj6z4+wpXzVTua0mIDe1Yks+ztOC6fkZUBdTpB39Htmf76ANr1aFVl3yrRNDi8FD4ZBGd2yH06exj1Ojy2AVp1qvs5gegL0az6dRUAvs6+vNzv5Rp6VE1UYBThfuEAJGQksDHlZxj7nhQpBNj7kVyIVzQKLfxciBgXBMjv7fbvTigpECsimvOHGx4ersXFxdnajFqTeDWRaWunoaHh6+zLmslrcLGvOdyzLpSUlDB//nzy8/PR6XTMmTMHL6+aF4E1o8aWbxNJipHrFo6udrTt6kXG+Tx0ekG7Hq0IHRXI1dQ8dixJIu/a9TDg1p08Gf5wV1q1cauf0QXXpFTH8VXX9/l0g/u/gNa963dOoKC0gPtX38+FPBlm+5+o/xDVLqqGXtVz/Opxpq+djoZGa9fWrJ60Gqcjy2H187KBbw94ajvYOTToOoraYTAYWf52nLnexahHuhESWbdpykai7kPiJoYaWTQi3Vp14/4ucmE2vTCdBQkLLH4NBwcHhg4dCoDRaGTHjh216id0gqjZ15Vqi/PL+PXgFbKvFJKZVsDhLedZ9H97WffJEbOjcHSxI2pmCJN/17f+juLkJlkP2+woBAyaI2+4DXAUAB8e+tDsKO4NurfBjgKge6vujO80HoBL+ZdYlLgI+syEjqZzpx+D3e83+DqK2qHXSymQ8tnJ3T+coiCn6Yhl3k4oZ9HIzOkzBzd7eWP99ti35puZJQkPD8fDQ9YBOHz4MFeu1E4lVa/XMerR65LQN1Ou5QQQPMCPGXMH0n1Im7ovYAMU58GaF2DxgzIbGsCzHTy6Fu75O9jXbQH6Zg5fOcx3id/J0zp68lr/1xp0vhuZ02cOTnpp34KEBVwtugbj54O9q2yw819w+bjFrqeoHr8OHvSOkgXBivPL2L1cSYFYA+UsakteOmz9u5TA/qg/rP5/kJ5Yc7+baOXciqdDnwagxFjCe3HvWdpS7OzsGDFiBCDzC7Zv317rvqmJmRWcQmUMmx7M3Y/1wMWjnlMt52LgsyEVS5X2mQnP7IGgIfU75w2UGkqZGz0XzVQd79WIV/F2rkciYBX4u/qbM7/zS/P5JP4T8GoPd70hGxhLYdVzYLCtuu+dRP8JHcyS9qdiLzf7+vNNEeUsasPVX+GzobDzXbicABlJcPBbmVGctLHOp5sRMoP2HlKjf/PZzWb5CUsSGhpKy5YtATh27BiXLl2qVb/aFJdx8aynkygrgV/mwtejIdOUeeviDdMXw8SPwakO1fCqYUHCApKzZLnZQW0GMb7j+Br7pKamsmLFCj766CO+/PJL9uzZQ3Fx1fIsj/d83OyAfjj1A8mZyRDxJAQOlA0uHoR9Hzf8zShqhYOTHcMfvq4yvGNxks2l+G83lLOoDavnQF4lCWuGElj5FJTk1+l09nr7CrpR7+x/B4PRsho3er2+goT51q1ba9WvNnpNbl71mCK6fAy+HAm7P4DysOGuY2U97JCxdT9fFSRnJvNFwhcAONs583rk64gawm0PHDjAggULSEhIICMjgwsXLrB582YWLFhAfn7lf1sXexfm9JkDSGXh9w68J+tmTPwI9KbPcNvbkJFcaX+F5WnfoxXB/f0AyMssJmbVaRtbZH2EEOuFEJXXBb7e5k+WuJbVnYUQQi+EOCSEWGva7iCEiBFCnBJCLBVCOJj2O5q2k03Hg6xtW63ISIaz1SS4FWXLmgp1ZGjAUAa3kWGuSZlJ/Jj8Y30trJIePXrg5yf/eU6dOsW5c+dq7NOhtzdOblXrJbUKcMW3fR2ysY0G2DMfvhghR2UADu4w8ROY/h241bHQUTUYjAbeiH6DMqN8onyh7wsEuAVU2ycrK4t16yoPY75y5QqbN2+usu/EThMJ9pLCdrsv7Cb6QjR4d4GoP8oGZUXyQcOososbiyEPdsHJVX5/j2xPJe1Mto0tsg5CotM0bYymaTUlRjUPZwG8ANw4uf8O8IGmaV2ATOA3pv2/ATI1TesMfGBqZ3uyaiGyt/ZF+HqMlKw+slw6mBpuEEIIXo14Fb2QWkkfHvyQnJIcS1hsRqfT3VJ+taZQaTsHPaMeqXyR28FZz8jZ3Wp8UjeTmQLfjIPNr8tRGED7IXJtos/D9Uqwq47vT3zPkQyZHd/bpzfTu06vsU98fPwtJWlvJCEhocrpKL1Ozyvhr5i33417V44QI+fISn0A56Ih7qs6vAtFQ3B2d2DIg1IKBA22LTyBoazpOmshxDtCiGdv2J4rhHhDCLFFCHFQCJEghJhoOhYkhEgUQnwCHAQChRApQghv0/GZQoj9Qoh4IcTnpgf1eYCzad93Qoi/CSFeuOF6bwkh/l9tbLWqsxBCtAXGAgtM2wIYCfxgavItMMn0eqJpG9PxUaLWdyUr4lH9kykgb4Rn98ikrB+fgI/6wTvt5Y1y01/g6Aq57nHTjbpji448FPIQAJnFmXx++PPKzt4ggoODCQiQ7yElJYXTp2semgf18ubBP4bTdYA/bl6OeHg70Wt4AFP/1B/f9rVYV9A0OPg/+HSwvFmCnJq55y14ZI1cDLYwqbmp/OfQfwCZbf3XQX9Fr6tZtDAjo/qFUIPBQF5eXpXHI9tEMqztMACSs5JZmbwS9HZyOsokbc4vcyGr5lGdwjIED/AnsJvMLbp2MZ9Dm5r0Z78EmHbD9lTga2Cypml9gSjgvRvuhV2B/2ma1kfTNPOTrBCim+k8gzVNCwMMwMOapr0GFGqaFqZp2sPAV8Ajpj46YDrwXW0MrWMNyzrzb+BVoHzeohWQpWla+cpTKlB+Nw4AzgNomlYmhMg2tbdtWINvCAT0qzozV2cHvt0g/YSMgimnOAdSdsmfcpw85RNnmz7QRv5+uvfTrD29lqziLBYnLmZK8BQ6eHawmPlCCEaNGsX//ifl0bds2ULHjh1rHB14t3Xnrsdqp8pagbx0GSl2csP1ff69YPIX4FeP89UCTdP4696/Ulgm5R6e7P0knVpUn/GdlZXFvn37SEysPqJNp9Ph6upabZvf9fsdey7swaAZ+OjQR9zX4T5c/XvBkJdh5z+hJE9Kmc9cYfHRlOJWhBCMeDiE7/8aQ1mJkdj1Z+jU1wcv/+r/jrZA07RDQghfIUQbwAc523IJ+EAIMQwwIu+NfqYuZzVNq0xWehTQD4g1/W87A7dIXGualiKEuCqE6GM65yFN067WxlarjSyEEOOAdE3TbrzLVvafotXi2I3nfUoIESeEiKtt/kCDGT8fnCpZQxI6mPy51C360wWZSDbuA1lNzb/39SfLcoqypZTFnn/D8kdhfiie80N5vli2K9PKeHfv324ZgTSUjh070qGDdEAXL14kKck6NcFJXCMT7ModhdDB0Ffgia1WcxQgxRn3XtoLQOcWnXmi5xNVtr106RIrVqxg/vz57Nu3D4Oh+sACHx8fs95WVXRs0ZEpwVKy/GrRVb5KME07DXsFfEy1o3/dAoe/r+U7UjQUD29nBkzoCICxTGP7d0lWrwrZAH4ApiBHBkuAh5GOo59plHAZKP8SVhVNI4BvTSOIME3TumqaNreKtguAR4HHgP/W1kiryX0IIf4BzALKkG/UA1gJ3Av4m0YPkcBcTdPuFUL8bHq9VwhhB6QBPlo1Bjaq3EdmCkR/BEkb5LRT+0iIfB4C+1fdp7RIRgFdPAiX4uFivMzN0CreoMqAqQH+nHKQIamfZBYx1Lu3HIGUj0Q82jToqfT8+fN89ZW8ifn6+vL000+by7E2mKJs2PAaHF58fV/LjtKRVvf5WICMwgwm/jSRnJIcdELHovsW0cunV4U2mqaRnJxMdHQ0Z85ULJbj7OyMn58fKSkpVV5j7Nix5nohVXGt6BpjfxxLXmkejnpH1k5ei7+rP6TGSZl1zShHls/tB3f/er9fRe0xGoz88M4Bc72LVgGu6O10ePo402NoAAFda1cLxUJU+c8rhOgBfAl4A8ORU1GdNU2bI4SIArYC5dMNazVN63lD3xQgHPAFViGnodKFEC0Bd03TzgohMgFfTdNKTX0cgATAHuiiaVqtQjEbRRtKCDECeEXTtHFCiOXACk3TlgghPgOOaJr2iRDiOaCXpmlPCyGmA/drmja1uvM2N20oQMpYpx2V0tvlPxlJxDja80RrOdIMKinlxwuXqBCT5OpbYfqKNn3qfNNZvHgxJ0+eBOD++++nd++GyWkAcGYn/PSsLAJUTvhv4J6/gYP1h/2/2/47Np3dBMCs7rMqhCSXlZVx9OhRoqOjSU+vOCL38vIiMjKSsLAwHBwcOHnyJPv27ePSpUs4Ojri4+PDqVPXM4EnT55MaGhotbb89+h/+eCArKI3ruM4/jH0H/LAz3+W61kAIeNg2iI1HdVIpKdks3xe5VPI4WOCzKOPRqDaP7gQIgHI0DQtyrRgvQZ5M48HBgP3mZpW6iw0TcsQQkwD/oicMSoFntM0bZ8Q4h1gAnDQtG6B6d6bZVrTqN0bsIGz6IgcarUEDgEzNU0rFkI4AQuBPsA1YLqmadWuxjZLZ1EZJfmQlsCLsW+zJU8++b56NZNZOTVIjLv5X3cc5Y7EzbfK5mlpaXz22WcAtGzZkueee85cjrXOlBbClr9VTDxz85fJdV3uqt8568jWc1t5YZsM7AhwC+DHCT/iYu9CYWEhBw4cqFSmPSAggMGDBxMSElLjyComJoYNG+SUmhCCadOmERISUmX7YkMxE1ZO4GL+RQCWjF1CD+8esnb4p4OuJyI++A30mFzPd62oC4e3nmf3sqrlPya93IeA4EYZYTSZpwPTwvZB4EFN02qtjaJUZ5sQ53PPM/GniZQaS3G3d2NtxFxaZiRfH4FcrUWCl0dAxemrNmHgel3q4ocffuDo0aMAjB8/nn79+tXd0IvxsPK3cOXE9X09JsPY98GlZd3PVw9ySnKY/NNk0gvliOGLu7+gm0s39u3bx8GDBykpqSgm17VrVwYNGkS7du1qH/oL7Ny505zQqNfrmTFjBp06Vb14vvHMRn6/8/cA9PXtyzejv5HXS9kN35iSD1195HRUI31WdzKL34wh81LVSbNdIvy45zc9GsOUJuEshBDdgbXASk3TfleXvtaOhlLUgUD3QGZ3n81XR78itzSPjzP283+R/3e9QVE2XDpy3XlciodrNw2+ci7InxNrr+/zDDRPX43o1JVjxwSaprFjxw569+5tLsdagdw0iPta3uR0euhyN4Q+JPftmAemxDecPKWT6DXF8h9INbwf977ZUUzyncTFvRfZdHRThTwSvV5PWFgYkZGReHvXTxtq6NChFBUVER0djcFgYMmSJcyePZvAwMBK298bdC8LExdy5MoRDqYfZOu5rYxqP0pqXoU/DnH/hfwrsPE1KcGusCrZ6dWXF67p+O2GpmnHgXrNvamRRRMjvzSfcSvHkVGYgU7oWDZuGV1bdq26Q2EmXDosn/bLnUgNiYSruJtDyGnPe4OdiBw4UI5EnE0RXxcOwML7oeimxFC9w/XkOoBOI+W0k0eb+rzVehObFsvjGx/Hr9CPHrk98CqoOI3g7OxM//79iYiIwM2tntLpN6BpGuvWraP8u+bk5MQjjzxC69aV102IT49n1oZZgKz3/dPEn7DX20NRDnwSCTmmEqAzlkPwPQ22T1E13/5xD3mZVWt8BfVqxdjnql+LshBNYmTREJSzaIKsSl7FX/b8BYD+/v1ZcM+COk2dUHDNFH1Vvoh+GLKvJyZl4c6HPIoBO1wo4AX+iyOlMoLJPxROb7vVUdyInbNcwI54otEXavOK83h24bO0utwKz1LPCsduXrS2JEajkZUrV5KQICVLXFxcePzxx6scsbyy4xV+TvkZkKq3s7pL58GpzfCdaRTmEQDP7pWjM4VViFl9mrj1KVUev/fJnnTuV/U6nwVRzsKW3K7OwqgZeXjdwxy9KtcWPhjxAXe1b+CicX6GHH1cOgQX41l/GvaXdAFgJLsZRh2Ub2csg+B7G2ZPHSlftN6yawtaccXvbNu2bRk0aFCtFq0bgsFgYNmyZeY8FQ8PDx5//HFatLg1Byc1N5UJP02g1FiKh4MH6+9fj2d5+dUffwtHlsjX/R6VeTwKq1BSWMaP7x3kauqtWfgd+/gw+sme9avHUneUs7Alt6uzgIpTGQFuAayatApHfc2KsLUlNzeX/8yfT2lZGY56eDHoJM6XD1SurnszUxdC9wkWs6U6yjOtb1601tDo2LkjUcOiCAwMrNvIqwGUlpayePFic75Gy5Yteeyxx3B3v1Vc8f249/n62NcAzOw2kz/0/4M8UHANPu4v1y5ASqB0GNYo9t+JlBSWcWjzOU7svURBbgmePi70HNaGnsMC0OkbTXhbOQtbcjs7C4DXdr3GutNSEfWFvi/wRK+qM5Prwy+//MLu3bsBuZA7atQo2PcZbPxD9R0f2yiTEq3IpUuXiI6O5ujRoxUWrQ3CwFm3s4weMZoZ/WZY1YaqKC4uZuHChaSmyrUHX19fHn30UVxcKtZTzynJYdyP48gszsRO2PHTpJ/MdUw49hMslwWU8AqCZ6IbJS9FYTPq5SyCXlvXF4hAZm6vT5k39lqDDRHiv0C5wkbPmtqXo+pZNGFe7PsiznbOAHxx5AvSC26RemkQgwYNwtFRjlb27dsnBfPCZoBDNYvCXkEQOMCidpSjaRqnTp3i22+/5fPPPychIcHsKHQOOo63OM6GwA049HTgob4PWcWG2uDo6MjDDz9sln9PT0/nu+++u0Wd1sPBg2fCngGklEt5wh4APSZBN9PoLDMFtr7VGKYrmglBr63zD3pt3XbgAPAZMgftQtBr6/4U9Nq6ho5SvgFG17WTchZNGH9Xfx7v+TgAhWWFzD9o2bltFxcXBg0aBMjplV27dslqdWPerbyD3gHG/VsW+bEgZWVlHDp0iE8//ZTvvvuugiSHl5cXA0YOYG3AWhK9EhEOgjci32i0aaeqcHZ2ZtasWbRq1QqACxcu8P3331NaWlqh3ZTgKQR5BAGw5dwW4tJuGAmP+dd1zbF9n8B5y1dMVDQ/gl5bpwc2IKU/bsQJeAt4piHn1zRtJzLxuU4oZ9HEebTHo7R2lSGaq39dTcKVBIuef+DAgebpk7i4OLKysuToYtZPch5d6KWTCBkHj/8MnaJqOGPtKSwsZPfu3cyfP59Vq1ZVkORo27YtU6dO5bnnn2NJ/hIKkYqyz4U9R6BH5TkOjY2bmxuzZs3C01MuXKekpLBs2bIK4oT2Ovtbal4YyysFuvvB6HmmI5qs211WdZin4o5hLBBWzfE/Br22rtFz5JSzaOI42TnxcvjL5u15++ddv9lYAEdHR4YOHQrIaJ+dO3fKA52i5MLr61fhL+myql1AX4tcMysri40bN/LBBx/wyy+/VJDkCAkJ4fHHH+eJJ56ge/fu/HDqBw6mHwSgR6sezOw+0yI2WIoWLVowe/Zss4z5qVOn+PHHHysUVBrWdhgD/OXU3fGrx83rUACETofOpki3jCRZ511xpzOqhuNtgeDGMORGlLNoBtzb/l76+sob9ZGMIxVvNhYgPDzcHM1z6NAhrl69Qd5eCIvlUly8eLGCPHh5dJOdnR39+vXj+eefZ/r06bRr1w6AtPw08zy/nbDjzUFvYnez7HsToFWrVsyaNcssZX7s2DHWrl1rXm8RQvBKxCsI0xrn/IPzzbU3EEJO7ZWvE+16X2bpK+5kavM02Ojl/5SzaAYIIXit/2vmm82/D/ybglLLyRTY29szfLicHtU0jW3btlns3DcuWn/xxRcVFq2dnZ0ZPnw4L774IuPHj6+Q4KZpGn/b9zcKyuT7fKznY9VnstsYf39/Zs6caZZOOXjwIJs2XZcfCWkZwsTOEwG4XHCZhccXXu/cIhDuflO+1gxyOspQce1DcUexqYbjKUCtBQAthXIWzYRurbpxf5f7AUgvTOero5at69ynTx+8vKRsxtGjR0lLq0W+RTXUtGg9duxYXnrpJaKioiqV5Fh/Zj07U+WUWAfPDvw29LcNsqcxaNu2LTNmzDAr+e7du5cdO3aYj8/pM8cc3bYgYQEZhTcUgez3OLQfLF+nHYHo/zSa3Yomx89AZdXwyvlryryxtapBURlCiO+BvUBXIUSqEOI3temnnEUz4vk+z+NmL2+s3xz9hgt5Fyx2br1ez4gRI8zb9R1d1GbRes6cOURERFQpyZFZlMk7+98BQCB4c9CbFk1ItCYdOnRg6tSp5kzy7du3s3evrOLn6+LLYz0eA2R020eHPrreUaeDCR9KKRWA7e/AlZONaruiaZAyb6wRmQex9qZDOcALKfPGft2Q82ua9pCmaa01TbPXNK2tpmm1evJUzqIZ4e3szW97yyfsEmMJ78e9b9Hz9+rVCx8fHwCSkpLMSWe1oS6L1jVJcrwT+w6ZxZkATOs6jT6+ferxbmxH165dmTz5er2Kn3/+mYMH5SL9Iz0ewddZahGtTF7JycwbHEKrTjDyz/K1oRhWPw/Gej9AKpoxKfPGXk2ZN3Y80BVZZnUS0CZl3libDTlVBnczo9RQyuTVkzmbI5Vl/3vvf4nwr77kZ11ITExk6dKlgKzdPXv27GrbX7x4kejoaI4dO1Yh09rOzo7Q0NA6y4PvTN3Jc1ueA2SeyU8Tf8LVvnlmNh84cIA1a9YAct3pgQceoGfPnvyU/BP/t0dKz0e2juTzuz+/njdiNMgyrBdM1d1GvwMDn7aF+QrLouQ+bMmd6CwAdpzfwfNbnwegq1dXlo5bil5Xz4p3N6FpGl9++SUXL8pqbx06dMDDw4MuXbrQrVs39Hp9jTWt6ysPnl+az6RVk0jLl+sln4z6hKFth1rkfdmK6OhoNm2S65U6nY7p06fTuUtnpq+dTuK1RKCS95meCJ8NBWMp2LtIZVqvIBtYr7AgylnYkjvVWWiaxtO/PE30xWgA3oh8gynBlis+dOjQIVatWnXLfj8/P/r27cuBAwduqWndsmVLIiMjCQ0Nrbc8+Fv73mJJklRjHdtxLPOGzquhR/Ng27Zt5oVuOzs7Zs6cSbpjOr/ZJNcVO3l24ocJP1QMC97+Dmx/W77uMBxmr1J1u5s3zf6Pp9YsmiFCCF6NeBW9kKOJDw99SE5JjkXObTQaiY6OrvTY5cuX2bBhwy2L1tOmTeP555+vdtG6Jg5ePmh2FF6OXvwhogYxw2bEiBEjGDBAJuWVlZWxePFiAowBjAgcAcCv2b/y46kfK3Ya8hL4msp9ntkBB//XiBYrFLeinEUzpVOLTkwPmQ7AtaJrfH74c4ucNyUlhStXrtTY7sZF627dujWojkSxoZg3ot8wb7/W/zW8nLyq6dG8EEJw7733EhYmFRxKSkpYtGgRj7V/DDshRxMfx39MXskNNRfsHGDiRyBMn+umv0DOxcY2XaEwo5xFM+aZ0GfMBXUWJy7mTPaZGnrUTG3yK8aMGVMh07qhfH74c1JyUgApjXFfh/ssct6mhE6nY8KECXTv3h2QIcZbV27lwbYPAtLhL0hYULFTQF8YNEe+Ls6BtS9DM542VtSDuZ59mev5W+Z6zmSuZ0tLnFIIESiE2CaESBRCHBNCvFCbfspZNGM8HT15PkwudJdpZUR9LrUAACAASURBVPwr7l8NPme5ZEV11CW6qSaSriXx9VEZNu5q78r/Dfw/myvKWgudTsf9999P586dAcjLy8Mx3hFv5Oe58PhCLubdNHoY8UdoJdtzcgMcXdGYJitsxVxPf+Z6bucmiXLmev6JuZ4N/QcpA36naVo3YCDwnBCie02dlLNo5kwJnkLnFvJmsjN1J7sv7G7Q+bp27WrOQK4MNzc32rdv36BrlFNmLOP16Ncp08oAeKnvS/i7+lvk3E0VOzs7pk6dav4Mc3NyuSvjLhwNjpQYS/j3wX9X7GDvDBM+wrw+uuFVWSJXcfsy19PaEuWXNE07aHqdCyQCATX1U86imWOns7terhP4Z+w/KTXWX1fI1dWVqKiqZchHjx5drTOpC4uOL+L41eMA9PXty4NdH7TIeZs6Dg4OPPTQQ7Rp0waA0txSotKjsDfYs+HMBo5cuUlIsH0k9H9Svi64Kh2G4namRoly5npaRFFTCBEE9AFiamqrnMVtwMDWAxkZOBKAM9lnWHpiaYPON2TIECZNmmQu7APQpk0bZsyYQc+eta7CWC3ncs7xUbyUu3DQOTB30Fx04s75Ojo5OTFz5kxzxrxrkSuDLg9Cb9Tzbuy73BLSPuoN8DStER1dAScsqzysaFI0ikS5EMINWAG8qGlajeGUTU/vWVEvXgl/hV0XdlFqLOWTw58wtuPYBkUUhYWFERoaSl5eHjqdzlyvwRJomsabe9+k2CAL/TwT9gwdPDtY7PzNBRcXF2bNmsXXX39NZmYm3sXeRF6OJFpEs/nsZu4Juud6Y0c3GP9vWCTFJFn7shQedG5hG+MV1sTqEuVCCHuko/hO07Qfa2oPamRx2xDoEcis7rMAyC3J5eP4jxt8TiEE7u7uFnUUAD+e+pH9afsBKd39SI9HLHr+5oSHhwezZ8821xPxK/Kjf3p/3o97nxJDScXGnUdBmKn4U14abPpzI1uraCSsKlEuZATJV0Cipmm1FphTzuI24qneT+HtLCNrlp9cTtK1JBtbdCvpBem8F/ceAHqh581Bb2Kvs7exVbbFy8uL2bNnm8vbBhQE0DqlNYsTF9/a+N6/g5uffH1oEfy6tREtVTQSNUqUMze7IQqTg4FZwEghRLzpZ0xNnZSzuI1wtXflhb4yZNqoGfln7D9vnfu2IZqm8da+t8gtlYq0s3vMpnurGiP27gh8fHxk8SQH6Tjb57Vn5+adXCu8VrGhsxeM++D69uoXoDgPxW3E3OxqJcqZm91QifLdmqYJTdN6a5oWZvpZX1M/5SxuMyZ0mkCPVlImYn/afraeazpPnpvPbmbreWlPO/d2PBv6rI0talq0adOGWTNngSnYrF12Oz5eXsl0YshY6GFau8g+B1v+2nhGKhqHudlXmZt9i0Q5c7OVRHl9uFOFBGsiPj2eWRvk+kWAWwCrJq2yefGg7OJsJv40katFsr63paXVbyeOnjjKsqXL0GnyWa7v4L5MuHtCxUb5GfBRBBReAwQ8tkGG2CqaKs0+01SNLG5DwnzDGNNBTkFeyLtQsd6zjXg39l2zo5gSPEU5imroGdKTdoPboSEf5A7uOcj+/fsrNnL1hvv+adrQZKGk0sLGNVRxR1GlsxBC9K3upzGNVNSdl/q9hJNeSnd8ceQL0gvSa+hhPaIvRrPqVyl57uvsy8v9XraZLVVRVlrKiT072LHov/x/9s47PKoq/eOfMzWT3glJCCRAQgtFUIqFYgEVFZDmWlB2FRS7/hRXV7GuDVfX3kXFBUEUxIIoRRAQpNcE0hPSSW/Tzu+PGwYCmUkhmSRwP8+Th8zcMmfCzP3ec973/b5bvllEYVZGm47n1jG3kh6V7nj8448/snv37ro7xU+G2HHK74VHYN3ZYemu0j5xNbOYX/vzNkp13wfAh7W/q93k2zlhXmHMjJ8JKP2e39jxRpOOt9slvyfm88qqQ7y2OpEd6UXNCpZXWip5ZvOJNfUnhj2Bj8GnyedpTfLTUvj0vtns+fh7zBsLObbqCEsf+Se/ffIu0n5G6ezNRqfRMfOKmewM2ul47rvvvuPgwYMndhJCCXYbfZXHm96EoztRUWkNnIqFlHK0lHI0kAacJ6UcIqUcjFIafsRdA1RpPrf2vZXOXp0BWJG0gr35ext1XH5ZDRPf+YNbPtnK22uT+O9vh5n0ziZmfraNSrO1SWN4c+ebZJVnATC221hGRzm3EmkLzNVV/PLSG1ziOZERodfR238Y/QNHcnWXO9Bsq2HbikbVK7UKF0VcRKe4TuwL2Aco2WRLly4lKSnpxE6+4XDFs8rv0gbL7waruZ6zqaicGY2JWfSSUjquMlLKfbj2LVFpJ5h0pjpLPi9ue7HB2YGUkrsWbmd3Zslp29Ym5PPEt/sa/fp78vew8OBCAHwNvsy9YG6jj3UXh35fz/meV+Cl86vzvJSSnn6DKV6bit12JintzUcIwcPnP8zhgMMk+Ck1MzabjUWLFpGefmKJivNmQPQlyu+5++CP1+s5m0pHJH5B/HnxC+JnxS+Ivyl+QXxLWZR7CCG2CiF211qUP92Y4xojFgeFEB8JIUYJIUYKIT5EcSlU6QCM7TaW80KVENOe/D38kOLaU2hnRjHbUoucbl+++yg5JdUNvq7FZuGpTU85grSPXvCoo2CwPSClpCg7i4L1h/HUnb4sdtwmPUbfj6IjWUhr2yxHxQbEMrHHRPYF7CPJR5lRWCwWFi5cSHZ29vHBwjX/Vfp1A6x/WenjrdJhiV8QHxa/IH4dp1iUxy+I/2f8gvgzzayqAcZIKQeg3PiPE0IMa+igxnhD3YZiiXu8QcbvwLvNHaWKexFC8OgFjzJ95XQkkv9s/w9juozB8/iF5RR2phe7PJ/NLvn6rwxmj+yOQef8XuOjvR9xpFhZrRwRPoJrYq5p/ptoIarKSknft5u0PTtJ27uL8vxCLuo0Cer/UwDgofOi6tM0skhDGLVovfVovPRovA1ovfRoah87nvfSo/U2oPHSIbQtk2x496C7+THlR3YF7cIDDyLKIqipqeGLL77gtttuU8wIA6Ph0ifh57lgt8DyOfD31aBpGYdgFfcRvyD+uEX5qSs4xy3Ki4F3mnt+qSwvHK/k1Nf+NBiQVOsszhGe2vSUo8/zrP6zuHvQ3fXu9+WWNJ74ruGlJm+jjgt7BDEqLpRRcSF09jM5th0pOsKUlVOw2q2YdCa+ve5bIrwbtMtvcawWC0cTDjjEITclCYPwINwzhnDPnoSZuqHXtF79iTDp6hWUE88ZToiMpx6hdX7D+P7u93lr11sIKZhYMRGRr+zr4+PDzJkzCQgIALsNPhkHmbVptlc8d6LT3inYzTaq9hdiLahC62PAFB+M1uvctl1pZRo9G4hfEH8tsNzFLplA9N4Ze5sWQDx5MEJoUWYtPYC3pZQNNr13KhZCiL24UBspZf9mjrPFUMWi8RRUFTD+2/FUWCowao0sn7C83gv40eJKLnxxbcO3GafQK8yHkXEhXNIziLcTHmBvgdKTYe4Fc7mx940t8A4aRkpJQXqqQxwyD+7Haq7BRx9IuGcPIjx7EGSMqNcKXUpZb4e+SmsZoouBgIBw7BUW7BUWbOVmZHULxzEEaEy6k4TFcNJMRY/VJHl8x5OkWtMp01Ryg7iJ3IxcQPGWmjlzpmJGmJ8A710ENjPoPODOTRDUvc5LVR8u4tj/DmGvPOlao9MQcG13vC44u5tPtSFNEYs3gHsb2K3v3hl7D5zZkEAI4Q98C9xTG492iqtlqPFnOhCV9kOwKZhZ/Wfx2vbXqLHV8NpfrzF/1Pw6+0gpWbA5zaVQ3DwsCr1Wy7rEPJLzKxzPH8op41BOGZ/s+RyPMEUoIky9GNV5Qmu8HQdlxwpI33tiaamypBiBINgjkn7eIwj37IGPvv64oC7YhEevQKqTi7AerTxtu13a2Zr/A7kZaYy8+e8MuWWiY5u02muFo1ZAKizYyy3YK8yO507eLmsaEBcJ9kpr7QW8/uK6x7jV8XsNFn7xqCGXYoqKivj0zQ+ZHDcWbz9vNN3mo038Co29FM3SeWhv/AjhaUBoBNbCKgo/P4C0nBKDsdopWnYYbYARj57Nt7ZXaRFa3aL8OFLKYiHEOmAc4FIsGrUMJYToCvSUUv4qhDAButp2fG2KOrNoGmabmYnLJ5JepmTSfDr2U4aEDXFsf+2XBP67RokzaAR46LVUmpWLXCdfI/de2pMbh55oqZpeWMm6xDzWJeSzKamAGgrwivkPQmNBSi2VyfdiN3eiV5iPY7lqcNcA9Gewlm+uriJj/17S9u4kbc8ujtUWz+mFgTDPGMI9e9DZFINRazr9YAGGrr6Yegfh0ScQfYgSrLCbbZT8kEzF9lywKt8HXaiJws6F/Pzdm47DL7phBkMnNK+bn7TYawXFfJKwWOr+Xm5WRKbccvrFvB5qsPCDYQfHNMryc4jdlyvNgzDUdw+oEWi8dGCTdWcUp2Ds6U/I3+Ob9R5VXNKUmcWVgCtjv1Sgx94Ze5s1vRVChACWWqEwoViivySlPNW4sO5xDYmFEOJ24A4gUErZXQjRE3hPSumym5MQwgMlGG5EmcEslVI+JYT4DKW37PHczFullLtqPdbfAK4CKmuf3+HqNVSxaDrrMtZxzxplHbtXYC8WXb0IrUbLm78dZv7qRAB0GsE7N57HJbEhHMwuRasR9O7s6/IiX2W2MuPH2zlYovx/1ORfhrngstP28zHquLBHMKN7hTAyNpQwPw+X47XbbOQkHXaIQ/bhQ45UVi+dH+GePQj37EGoRxc04vRgrjBo8YgLwKNXIB69Al2uy9srLVjyq9AYteg6eSKEYO+aX/jlgzeh9nsyYsqNDJ98g8sxtwR2s+0UQVGE5GhuJpuTNuJn9SGUIMK0nVhRs4USocyMwuz+jDMPREfzAttCryHi2Qtb8q20C0qrLfy8L4f8shq6BXlxWZ9QjDq3Bv+bIhYa4A/AWYbSzL0z9jbbeVYI0R9YgGJZqQG+llI26EbZGLHYBVwA/CmlHFT73F4ppcvbj9qLv5eUsry2K9NGlIyq2cBKKeXSU/a/CrgHRSyGAm9IKYe6eg1VLJqOlJJZq2exOXszAE8Nf4rC7EH8+6dDAGg1gjdvGMRV8Z2bdN7lR5bzxB9PANDDvwevjviMP44UsfZQHpuTC6l2cqd8fNYxOi6E87oGoNMIinOzSduzi7Q9O8nYv4eayhPLXUHGcMI9uxPu2RN/Q0i959T6G/HoHYipdxDGGD+Ei6ytxrB//W+sevcNpFTew7BJ0xgx9aZ6Yxzu4P619/Nb+m8A/HPoP7ky7Eo+/eQTSkqVzpgxnbsyYeAVsPdX7Cl7sEs/bJ4x2H36YMmrBJvz77ww6Yh46uwyJFzyVwZPrdjvmCUDhPgYeeuGQQyNCXJxZIvSpA9L/IL4IOAz6oYDSoF/7Z2xt00cNBojFn9KKYcKIXZKKQcJIXTAjqYEuIUQnihicWftT31i8T6wTkr5v9rHCcAoKWW2s/OqYtE8jhQdYfL3k7FJGyaNH3kHHwC7B0LA69MGct3ApmUuFVQVcN1311FqLkUg+PKqL+kfcuLjUW2x8WfKMdYlKEtWKQUVdY73sFUTWZ1FTE0WMeYs9FUn0ne1Qk8nU1ciPHvQ2bMHJm39Xfv0kd7K8lLvQPSdvVr8Qn7wj/X89NZ8h/3HkGsmccmNt7WJYKSVpjHhuwlYpZUAYwArJ63EUmbh008/pbxcWZLq27cv1183Hs0HI6GgtgnWxPcpK76Qkp9SnZ7b87xQAqfGueFduIdNRwq48aM/643DeRq0/PLAJUQGuMidbjma9UGJXxAfCwwBKoBf987YW9HAIa1GY8TiZZS83ltQ7vzvAg5IKRvs6VhfelbtMtRwlMKQ34C5UsoaIcRK4EUp5cbaY38DHpVS/nXKOe9AWRYjKipqcFpaWhPe7plRkpdD8s6/sFutRMT1IazHGfdMbzP+/ee/+eqQ0onNXHgxNXlX88rk/kwZ0qXJ53p4/cOsSl0FwM19buaR8x9xuX9yTjG/rf+TpF07kFmHCanJq/NN8tB6E+7ZnQjPHoSauqET9azB6zR49PDHo08gpl6BaH1b34I9cctGfvjvK45lsPOuvJZRM25vE8F4aetLfHnwSwBu63sbDw55kNzcXD799FOqq5WiyUGDBnHtoDDEJ2MBCR7+2P++mdxPMrEV1Zx+Ur2GTvcMQh/qlounW7j54z/ZcLjA6fZZI2N47Mre7hhKh7cob0xR3lzg78BeYBZK4OWjxpxcSmkDBh5PzxJC9AMeA3IAA4o54aPAM9T/xzxNyaSUH9Qex5AhQ9xSJGK32Vjz6Xvs/vVnx9o1QJe+/Rl//6N4+vq5OLp90llei7R+h9BVog/8g7vPv7lZQrEmfY1DKCK8I7h74On1G85SWk9eRPI3hCrBac9Ygo2d6n2tch1URHoTPiSMTv1D0RjcW3AWO+wiNFod3//nRew2Kzt+WoHNauXSmbMRGve6/c8eMJsVSSsoNZfy5cEvmRo3lchOkdx00018/vnnmM1mdu7cidE4jLFD70T8+Q5UF6NZ8yghd3xI0bLD1ByuW4BpiPA+q4Qir7SazcmFLvfZmnLM5XaVE7itKE8I8RRQIaV89aTnRgEPSynHt+dlqN+/+oxty5fWuy2ydz+mPvXvNlu/bg7LdmTy0JLd6Pw34xGm1P6MjBzJW5e+1aTzlJnLmPDdBPKqFPvzDy7/gOHhynp3+bFC0vbuqpPSejIatISaoojwjqWLTxxGWU/2EnAEG39g5Q+sHMTmuHvo09mXUXEhjIoL5bwof3QtVC3dGJJ3bmPF/BewWSwAxI+5gstvv9vtgvHFgS94eZvS02Jst7G8OlL5aqWkpLBw4UKsViXraeRFIxi9//+gKFU5cOrn0Oc6rMeqsRRUUrIiGWuBkq7b6aHBjiyxjobVZmdnRrFjuXP/0dIGj7mgWyBfz3ZLjKbjXCCc4Koo72sp5VRnxXkNxSycpWcB26WU2bUB8P8A1VLKuUKIq4G7ORHg/q+U8gJXr+EOsaiprOD92TOw1Dj3Q7rijnuI6N0Xg4cJg8mE3ujh9gtHY/l+91HuW7QTuwSw0a3/+xRalFTady97l4siLnLsW5CRRsaBvWg0WroNOA+/0Lp3/E9vfpqliYqIToi6htv8JjrEoTAznVMxaEyEe8YQHTyQYG04Gns9fyOtwBjjh6lXIKKnP1uLKlifkM+6hDxSC0+vhQDw8dBxSc8QRsaFMCo2hFBf1xlWLUHqnp0sf/lZrBbF4bXvyEu5Yva9aNxor2GxWZiwfIIjFfqLK79gYKjiEJGYmMiiRYuw18ZYrhjSnRF/1c76vEJhzp/gqdSfVGzPpWiJkgnnNTSMgIk93fYezpS8smrl85GYz4bEfEqrm1bUPPfKXswe2b3hHc+cs1osOtde1O9BCU7Xma9JKV0GC5ylZwkh1gAhKH+8XcDs2owpAbyFUhxSCdx2arziVNwhFun7drPk2QbDM3URAr3RA4PJ5BAQg4cJ/fHHJ/9uMp20n+eJbSdt1xmMLTJz+XlfNnO+2olNUQoeGRfHoNh87lh9BwDRftF8c+032Kst/PT2fJL++rPOe4ofcwWXzrwTrU7H1qN/8uiSuwgvMNH1mA8hRYZ63Vl99IFEBw8gyq83njXeiHq+M8Kkw9QrEI/egXjEBqDxqH91NKWgwnHXuCW5kBon5n59Ovsyupcy6xjUpfVmHen79vDty09jrVHW/3tdOJIr5zyIRus+wfgt7TfuX3c/AP1D+vPllV86Piv79u1j6dITM+JroioYnP6B8mDADTDxPUApMMx+aRv2MjPoNHSeez5ab4Pb3kNTsNrs7MooZl1CPmsT8pzOHgw6DcNigujb2ZdP/kip97MS4W/ix3svxs/TLTYnZ69YOHYQ4jlgOrAD+ARYJduJoZQ7xCLzwD4WP9221tpCaNB7nCI+JhN6D8/TBMlgOi5EnrX7eGDwMLEtq5K53ydQaddhE1ruvzyW+y9TAvT3rrmXtRlrAcWew/P7JJK3b613LJF94jF4mkjY/Sd6Sz0XfgRhvt2JjTifYBGBrqr+C6cu2ORIbzV09XXpi1QfVWYbW5ILFfFIzCfNyazD10PHxT1DGBWnzDxCfeqfdUgp2ZlRzP6sErw9dIyJ69Soi0jmwX0se/FpLNXKMk7s0Au56t7/Q6trTDjwzJFScuvPt7IjTylJeuWSVxgXPc6xffv27Xz//feOx5M9NtGvuvYmYPTjYAoA71BKs/tTujoTAN/LovC97ETxZVvT2NlDVKAno2uXJofFBGGqjWn9lXqMf367l8Tccse+w2ICefn6AUQFuW3JrVlicbBX7/OA81GyoX7sfehgiwRZapOP/gKypJSNcutobAW3AK5AcaAdAnwNfCylTHJ5YCvjDrGwms28f+cMqsudF6z3v+xK9EYD5qoq5af6xL+W44+rqxx3oG2NFBo8PE8IjTRo2V96kBqtDYPQE56j3FV66wIINUUhpZ2cqlSqbPX/DQxaD3pFj6CLbx+8K7zBXM9n6nj1dJ/a9NYWXhdPKahg7SFFOLYkF2J2MuvoG67EOkbHhTKwdtZxtLiKuxbuYFfGibiKh17Dg5fHcsclDS9RHE08yDcvPIW5ShGs7kOGMf7+R9Hp3WPMt79gP9N/mA5AuFc4KyauwKg9kR22efNmVq1SkhA0QjBdfkssKXXOYTdGkl3xLtIq0Hjp6Tz3fIS+bRxrbXbJrowi1h7KZ11iHvuynM8ehkYHMrrWHSA62HnKtJSSfVml5JdX0zXIi+4h3q35FuqjSWJxsFfvMGARSgHzcaqBZ4F/9z508Ixu2IUQD6Jcy31bVCxqTz4ARSzGAWtRqgtXSyld50m2Iu4KcG//4TvWfV5/AljssIu45oHGzTzsdhuW6uqTRKWyrqjUERplm6W62vGcpaqyjhAdD7C2NDqh54KQq+nidSLf3i7tJJXtYmfhr0gkdh8vgoK6MzLkCozHtPU61QijFo/YAGV5Kc519XRLcnzWsbZ2ySr9mPNZx4U9gvkr7Rj5ZfV3l3v5+v5MPb/hLLGcI4ksfeFf1FQoafAx553PNQ88hs7gnuWcxzY8xspkxa3hgcEPMLPfzDrb165dy/r16wHQYeVvfItEwzH88aaCnqRSbrmdcpty3fCf1APvC5pWmHkm5JfVsD5RiU1tOFxASVX9n+0ugSaHOAyLCcLT4J4ZXAvQaLE42Kv38bt+Z03m5vQ+dLDZFuVCiEiUEMHzwIMtJhZCiHuBGUABSsrsd1JKixBCAxyWUrolOlQf7hILKSXbV37LlmWLHdXEGq2OvqMuZcyts9x2QTgVm9V6ktDUFRJLdTWJGfks3nwEYalBLy30CTbSO9hQK1iVp81+jsccLgydSKRXbB0n1uO/F1YfpcZTEG6v/0LS0tXTZ4qUsjbWoaxx/5lyzOmsoz66Bnmy9qFRaDQNf9dzU5JY+vy/qC5T7oS79h/EdQ8/jt7Y+gH3nIocxn87nhpbDd56b36Y9AOBHicMFKWUrFq1ii1bthx/hpOvX55UcrV9G37meYAGXYiJTg8MRjTifTeH47OHdQn5rEvIZ2/W6Z0ZAQxaDUNjAh3eYjEuZg/tnKaIRaMsynsfOtgsi3IhxFLg34APtdmojTquEWLxDMqS02kBbSFEbyllm7XkcncFt6W6mqxD+7HZrIR1j8XLv/26c25PK+KWj/+kotbi4IYLonhhYj+X0/RlB5fy7Q+f8WLhXKeW3fWh7+KjBKj7BKEP82zXX+ZKs7U21qGIR8ax+h1eT+binsH0DfcjJtiLmBAvYkK8CfSq/wYhPz2VJc8+TlWpcvHr0rc/Ex95Er1H6wvGf3f8lw/3fgjAtLhpPDHsiTrbpZR88/Ic9lWF1nu8BhvXW73wsyreUEG39sXUq0U6eQLK7OH3RCX28HtivtPZQ2TAidnD8O4davbgiqaIRaMsynsfOthki3IhxHjgKinlXSeXLjTq2HYSq24Wqt1H/ezOKOamj/6krEa58Zg8OJKXr+/f4N2xtdLMt+9/zPDcfi73swgruh4+BMRHYOoVhNa3fWbONISUkru/2skPe52W8jjF31NPdLAXMcHexIR40b1WRKICPanIzWLJs49TUay0p43o1ZdJc5/CYGrdYGqFpYKrl11NYXUhWqFl2bXLiPGPqbPPr6/+nY3lzpfVYmUOl9Qo/UeMXQyEzHFpz+YSZfZQzPqEPNY2YvYwMlYJTncP6bCzB1c0RSz+A9zfwG69ex86eKjJgxDi38DNgBWl854vsExKeVODx6picXaxL6uEv324xZExMmFgOPOnDkRbj1DYq63UpJZSk1RMTXIJlqPljWiuCLsuzmb81VNbeuhtws/7spn9pXNzY61GOFKNG4NGQESAiV6mGmJ3fIGmUlmSCo6JZdoTz+Dh1bqB1aWJS3l689MAXBJ5CW9f+nad7R+99jSZpc7fjw4L/6jpglX2BSA0/G0Mg4dDnwkQ0HCGVEG5MntYm5DPhsP5FFc6nz2MigthVGwow7sH4WU8K2YPrmiKWDTKorz3oYNn1IFLnVmcwxzKKeWGD7ZQVPsFvTq+M29MH+ioM7DX1IpDcgk1ScVYspyLg0TWWxNh1liIfHwEBq/WX1ZxB1abnQnv/OE04+atvw1iZGwIqQWVJBeUk5RfQXJ+Ocn5FaQUVFBlcf599bWUMDF7Bb42JWUz3yOUgwNuoEtYEDEh3spsJNib6BAvvFvoYmmz25j8/WRH//OTq+oBPvrgPTKP5rg8h4e0EmGPpKstmB4cINxQ2yQrYjD0nagIh3+X2teT7M4sZl1tJtrerBLqu6QYtBouiA6srboPoXuI99k4e3BFU8SiQYvy3ocONtui3DEgVSzOTY7klTP9g80UlCtZPZf36cTbUwZgzyinJlmZOZgzy5z3A95CNAAAIABJREFU19JpMHbzxRjtx67tfxJzLKxewdjeI4nr/nFr674ZN1NQXsNDX+9mfWK+4zl/k565V/Zi+gVRTo+TUpJTWk3ycQEpqFB+Lygns6gKKcHHUsrEnBX4WZW043xDMN+Fjaf6lOZMoT5GRzzEERsJ9iYywNTkosJNWZuY9essAGIDYvl6/NdoayvLT86KagxCCrpxlDhxiDiSCUAR1cKAAazXXcT7BfEkVPnWe2yEv8lhyTLi3Jg9uKKpqbNOLcp7HzrYPi3K2zOqWCikFFQw7f3NlJTV0A8tEwN9uczbE2tmOThbQtEJjFG+GGP8MMb4Y4jyQeg0JBcnM/nb65mTM53LSoairW2iUyPMfBe4lsVhv7B66mr8jB3PPLEhkvLL2X+0FG+jlhHdg/E4gzqDaouNtMJKUgrKOZKSScW3b6GrUEztCvSBfNf5Gqq0Dccw9FpB1yAvJT4S4kX32hhJdLAXgV4Gp3fns3+dzR9ZfwDw9IinmdRzEgDl5eW88847VFaenk6s0Wjo1asXmZmZlJbWP9MKpYA4koglmQhy0ADb7LH8YBvGaobSLboHo2JDGd3rnJw9uKK5RXl1LMp7HzrYfi3K2zPnulhIi53Mffl8891BetRI+qDF4OwzqRUYonwwxvgrAhHli9Cffse68OBCXtz6IgBBFj/6VvXAjp3dnomU6ZTP6Vtj3mJkl5GnHavinPJjhXz97OMUHVWqpD1Dw/GfdC+p1VpS8itILqggrbACi4vGRKfiZ9I7ZiDKv8rMpGuQJ5nlKVz//fXYpZ1gUzA/TPwBT70iTlsPpLD0m2/wtJ2oaK7GwLDRY5kwcjBSShIS00hYuIFM8snX1C8cXlQQSzJxJBNDOnpsiKjh0G8S9L4WfOp3Dz5H6fCqqYpFB0Ja7ZjTy6hJLqY6qQRzeqnzrmcagaGLD8bufhhj/DBE+TbK0vtksXCGKhbNo6K4iCXPPu4wWQzoHMGUJ5/HJzAYUOInmUVVpBRUkORY1lLiI3llja/+F0IJIGtDllGoUZacroq8mfuH3INOIxj3+gYKK2oIERX4iGqq0JNj90WjEUwd0oUD2WXsySzmdmnkFoxUUMMCbTa5mgI6a0rRidM/czqsRJNOHMnEkowvldDtIug7QREO7/rTdc8hVLFoS852sZBWO+bMMmqSSpS4Q1oZOCkoswGGSG88ewRg7O6HoWvjxOFUkoqTmLB8gtPtRq2R36b8dlYuQ7mDytISlj73BPlpit2GX6cwpv7rBXxDXF9My2ustTMQJcieUiskKQUVddqFnozQluHV/VWEtgZp11OR9DBauz/WRmZ3BSFYgjcGBGUGQcaU7pwf7U9uZhqJiYkkJCRQUVH/qkg4OY5ZRxiFiOiLoO8k6H0NeAU36vXPMlSxaEvONrGQNjvmzNqAdFIJ5rRSpJPe1XYkh7CzAyslwR48dvsQfP1aJkPpwXUPsjptdb3bjndlU2k+VeVlfPP8v8hNVjKWfENCmfrkC/iFhjX5XHWC7CfNRI4H2fWBazGGKr5QluLzqM5uOOVZrxUM6RrI6F4hXJFaje6A4l0XeEMcngNOiJrdbufo0aMkJCSQkJBAXl5evefzpZS4WuHoRja6mAuVrKre1zhs0s8BVLFoSzq6WEibxJxVpqSyJpdgTi1Bmp2kKwnQh3sju3jzysGj/FRSTiXQL8KXhf8Yhp+p5XyXKi2VzNs8j59TfkbW5tZqhZYbet3AQ0MeQqc5p7NaWoTqinKWvfAU2UeU/tjeQcFMffIFAsLCW+41LDYO5xdx1/ppFJuVC7kt4z4qy517Phl1Grb/63JHKq8lp4Lc15U6FH2EN6F3D3QatC4qKnLMOFJTUx29NE7GgJnupBFHMj1FOl7dhynC0etqxQH37EUVi7ako4mFtEssR8tPLCullCKdLCEgQN/Z60RAOtqPErudGz7cwqEcJQ2zV5gPi+4Yhr9n61RQZ5RlsD13O1qhZXj4cIJN5+TyQatRU1nJshfncTRBcW3wCghkyr+eJyii6e1tXfFD8g/M3aCYXXbS9+HInptxdu26vE8nPrxlSJ3n8j/ZR02iUo0eckc8xhj/Bl+zurqapKQkEhISSExMdPQFPxmBnUiylVmHSCO4+yBEv4kQdxWYGn6NDkazxOLt2WvqWJTPeW9MS1mUpwJlKCvYVinlENdHqGLRJOxmG+bUUqRdYoj0brBBjLRLLNkVjmWlmpQSZI3zIi59mFdtQNofY7QvmpP6KZRUWbjxoy2O4rGeod4sumMYQd5GZ6dT6QCYq6v49qWnyTywDwBPP3+mPPEcwVHdWuw17NLOTT/exN6CvQCI3NsoPRZ32n46jWDxrOEM7lr3Dr/6cBEFHyvj8+gdSPCMvk16fZvNRkZGhmO56tix+q93gRQpAXJNOlHde6PtNxHirgSP+us4OhhNEou3Z69xaVE+570xZ2pRngoMkVIWNPoYVSwaRkpJ+cYsSn9LR1bXXuy1Aq/BnfC/prsjBVXaJZacCkeFdE1KKdJFm0ddJ088uiszB0O0n1ML77JqCzd/vNXRbyEm2ItFs4Y5beSj0rGw1FTz3SvPkb53FwAmH18mP/Ecod1iGjiy8ezM28ktP90CQJgpkpq0B0kvPGHLHuhl4IWJ8Yzrd3rcREpJ3hs7seQowewz7dNdUFDgEI6MjAzquwZ5UE1PUogTGfToHoNH/+sgbhwYfZr9um1Mo8Xi7dlrGrQon/PemGZblIMqFq1G2cYsSlYm17vN2NMfU69AqpNLMKeUYK90IQ6hnsqSUndlWakxrSsraqzc+ulWtqUqywBdgzxZfMdwwloomK3SPrCYa1gx/wVSd20HwMPLm8lPPEenmB4t9honJy48MuRRuurHknaskhBvI6PiQlwWIbZWn+7KykoOHz5MQkICRw4fxlxPjxYNNrqSRZwmnbjoSAIGXQOx48Dg1SJjcBNNEYtGWZTPeW9MsyzKAYQQKUARiuHP+1LKDxo8RhUL10iLnex//+lSBJyhCzE5KqSNMX5ofZoWW6gy27jts61sSVam7RH+Jr6ePZwIf1MDR6p0RKwWC9//59+OlrZGTy+u/+czdO55+pJRc8gozeDa5dditVvxM/rxw8QfGp0CLa12sl/ehr209fp0W61WUlNTlSD5gX2UlNfftCqUAmJFOnFdw4gYciWa2LFgcFt71ObSFLFolEX5nPfGNNmi3DEYIcKllEeFEKHAauAeKeXvLo9RxcI1Nckl5H+wp1H76oKPi4Pyo/Vtfjyh2mLj9s//YsNhZZbY2c+DxXcMd2fPYJU2wGa18MMbr3B46yYADCYTk+Y+TUSvPi1y/le3vcqCAwsAmNBjAmO7jSXYFExcQFyD1hyl6zIo/TkVaP0+3VJKcnNzlQD5/t1k5dUf5/CigliRQWxkIN3PvxxD73GgP/1mStrt5OzfSPmxXAIjexDUfVCrjd0JTRGLRlmUz3lvTJMtyutDCDEPKJdSvupyP1UsXFN9pJiCj/a63MejXxD+13RH59cyweYaq43ZX2xnbYJibBfqY2TxrOFEB7t32m0rL6d63z6EVotHfDwaNzTwUVE6IP701nwSNm8AQG/0YOLcp+jSJ/6Mz11SU8KVy66kzFy3n3pcQBxPDn+S/iH9nR5rr7SQ/eJWpNnu9j7dZWVlJCYkkLB3O8kZ2fXWpmqxEiOOEtfZh9gho/CNvxL0HmRs/4WVP60i13oi3hHtUcq1N8wkoKvr3i0tSFPEolEW5XPeG9Msi3IhhBegkVKW1f6+GnhGSvmzy+NUsXCNvdpK9vN/Oi2OAwiZ3R9jt5apaLbY7Ny1cAerD+QCEOxtYNEdw+gR6r7AnrRayX/jDY59uRBZpXSS0/r7EzR7FoEzZqjmcG7AbrPx87uvc3DDWgB0BiMTHvkXXeOdxTwbR0lNCVctu4pS8+l+T546T/539f9Oa5h0MsUrkijfdBRwf5/u45jNZlKSk0nYtYXEpFTKnbSi7yzyifTVsrPEByt6TrSSVf7111Qw656HMAU0vRiyGTRFLBq0KJ/z3phmW5QLIWKAb2sf6oCvpJTPN3Rc2zZI7gBoPHR4DXdeKGXo5ouha8uk9lltdu5ftMshFAGeer78x1C3CgVAzvPPU/jhRw6hALAVF5P34ksc++QTt47lXEWj1TLurvvpO+oyAKzmGr576RlSagPgzWVp4tJ6hQKg0lrJR3s/cnm894Xhjste+YYsZBMaQ7UUBoOBuF69uHb6rTz42JP84x//4OJBsXTyrDuWbBnCtpLAWqGAE9dr5d9iuxc7flzgvoE3kjnvjbGjWJOvPGVTKXDfmQgFgJQyWUo5oPanb2OEAtSZRaOQNknx8iNUbK3bNMYQ7UfQTb2dprw2BZtd8tDXu/hul3LX5uuh46vbh9Evwr0eTObMLJIuv5x6O9gAGm9vev6+Ho2nGjtxB9Ju59eP32HPr8oKgVan45oHH6P74Oa1O73lp1vYmbfT6XYfgw+bbtjk8hyFCw9StVeJpQXN6IOpd1CzxtIaFBcVkbBtDQn795BaIrE3cD8cZSxl5mOvuWNozS3Kq2NRPue9MW1mUa76NjQCoRUETOqJz8hIqg8dQ9okxmg/9JEt49dvt0se/WaPQyh8jDq++PtQtwsFQMkPK50KBYC9vJysh/8Pr4suxBgdjSE6Gl2nTurSVCshNBou+8cctDo9O3/+HpvVyor5LzD+vkfpOXREk89ntbvO6qswV7AiaQVXdL0CD139MSrviyMcYlG+IatdiYV/QABDr7ieoVdcT3V1NS+/+Bx2F5c5m/PV5XbBnPfGJAKJbT0OUMWiSeiCTHhfGNGi55RS8sTyfSzdrvQ58DJo+Wzm+Qzo4h67A3tlJZXbtlGxaRMVmzZTc/hwg8eUr1lD+Zo1jsfCZMLQrRvG6G4YuikCYujWDUN0NFrvDpUL3y4RQjD61jvQaLVs/+E77DYb37/+Ilfd8zC9RlzSpHMNCh3kqOSuDzt2Ht/4OC9tfYlru1/LlLgpxPjVjWEYo5SlV3NaqaMDoyGy/RXLeXh40M1USbKTTn4AXUNatyf62YS6DNWGSCl5+vsDfLYpFQCTXstnt53P0JjWu1OTNhvV+/cr4vDHJip37YJ6CqFaCl1ISB3xMER3wxgdjT4iAqFT71WagpSSjf9bwNblSwEQQsO4OQ/Q5+LRjT5HZlkmE5dPpNp2uleTMwZ3GszU2Klc1vUyDFqltqJqXwGFXx4EwDQghKAbejXhnbiPpI3L+OLX+lPf9Vi467Yb3JUR1eGn3qpYtBFSSl748SAfblD6Ghh1Gj659Xwu7NHyZn3mjAwq/tikCMSff2IvKal3P+Hpidf551OTmoolLa3efXwnTiB45kxqUlIwp6ZhTklx/NicnLde9HoMXbpgiI5WZiTRJ2Yk2oAAdVnLCVJKNi35ii3f/E95QgjGzrqXfqMvb/Q5tmZv5eHfH6aousjxnIfWg38O/SeDQgexNHEpy5OWU1xTXOe4AGMA1/W4jsmxk4nyjiJn/l/YCqtBA2GPnI/Ov32mVu/47m1+3JWN9aSFFC+qmDz2QqKHX+uuYXT4D7QqFm2AlJJXf0ng7bVJABi0Gj64ZTCj4lqmm5itpISKLX/WLi1twpKRUf+OGg0e8f3wGjEC7xEjMA0YgDAYsBYVkXXPvVSe8rf1GTuW8JdedFpvYS0qwpySqohHaooiKCmpmNPTmzR70fj5YexWV0AM0d0wdO2KxqgaJwJsWbaYPxZ/4Xh8+e130/+ycY0+vtpazZr0NaSXpRNiCuGyrpfVqeausdWwOm01SxKWsCNvx2nHD+08lFnmG+i8UZlpeF8cgf/VLedl1dJUFWVzYO1SystKCQzpRK/R09Gb3LoEpYpFW9JRxeKNXw/zn1+VmJVOI3jvpsFc1qf5/Yql2Uzlzl0Ocajevx/q6SUAoI+KwmvEcLxGjMBr6FC0fvUH0aWUVO3cSeXWraDV4n3JJXjENc92QlqtWI4exZxykoCkKqJiddIwp16EQB8RcdqSVksF2W3l5ZhTUtF4eWGI7tbuZzfbVnzD7wtPZFGOuW0Wg8Zd0+Kvk1ScxJLEJaxIWlGnmM9oN/DlkRfwtnmCURD+2DA0HurSohOa9WGaP218HYvyhxavbCmLcn/gI6AfSuHJTCnlZpfHqGLhXt5Zd4SXf1Ya3mg1grf/Nohx/ZpW2CSlpObwYYc4VG77q05NxMlo/PzwGjZMEYcLR2CIjDzj99CS2MorFOGoFQ9zSgo1qcoSl6ys3xuoPk4Psp8ItjcUZLdXV5P3yqsUf/MNsrbvgjEujk6PPoLXiKZnHLmTHT8uZ+2CDx2PR978d4aMn9gqr1VlrWJV6iqWJC5hT74SB5iRdy3TC5UZzeqe2+l6eX8uibxEbZB1Ok0Si/nTxru0KH9o8coztShfAGyQUn4khDAAnlLKYpfHqGLhPj7akMxzPyhBQY2A16cP4toBjeuMZsnLo3LzZkfWkjU/v/4d9Xo8Bw1yiINHnz4IrXssGVoSKSXWvLwTAnLSjMSSleV05lQfroLsaLVkzJpFxe8bTj9QqyXq44/xGta8mgZ3sWvVD/z2ybuOxxfdMIOhE6a06msmHEtgSeISNias591Dj6GXevJ0x5jZ40mCvIK5vuf1TOo5iTAvt1RHN4qyYwVUFhfjGxKKycftPTIaLRbzp41v0KL8ocUrm21RLoTwBXYDMbIJAqCKhZv4fHMqTy7fD4AQMH/KACad5/wu315ZSeVffzkC065SWo2xsQ5x8Bw8+KwvmLPX1GBJT2+RILsuOBhrdrbTXUwDBtBt8aIWGHXrsue3Vaz+8C1HjcyIKTcyfPINrf66lZZKDn2+kdDDinnfi+GfsN5P+U5qhIZLIi5hStwULgy/EK2mbW5aCjPTWfPp+6Tv262MS6sldthFjL71Djx93VbL1BSxaJRF+UOLVzbLolwIMRD4ADgADAC2A/dJKV0W/Kli4Qb+tzWdx5adyG1/6fp4pp0fVWefpqS06kJCHOLgNXw4upCQVh1/R6Klguwn0+P39ehDWyb5oDXZv/43Vr37BlIqs65hk6YxYupNrR57OblPd75fKbO6PE2Vre6yaGevzkyOnczEHhMJ8XTf57U4N4eF/3yA6vKy07YFRUbxt+fnY/Bwi+V/U8SiURblDy1e2SyLciHEEGALcKGU8k8hxBtAqZTyX66OUxcWW5ml2zP557cnhOLZCf0cQtGUlFbP84fgPWIEXiNGYOjRo90HX9sKXUAAuoAAPM+ra0HtLMhetXu3I07hjIa2txf6jrwUjVbLT2+/hrTb2bJsMVaLhUtuvK1VPy/6MC+MsQHUJBYRUuLLqknLWSXXsyRxCYlFSiJHdkU2b+58k3d3vcvoqNFMjp3MsM7D0IjWtafb+t3X9QoFKDOO/et+bZWkgDOkMWusZ1J7nglkSin/rH28FJjb0EGqWLQiy3dl8cjS3Q73jKdHd+G6kgSyn/q04ZTWfv0cWUueAwciDC3baOZcQ+h0GKKiMERF4T3yRMywaNFicubNc36clxe6zu53Vm0uvS8ahUar48c3X8Fus/HX98uwW62MmnF7qwqGz8UR1CQqdRu2zceYfut0psVNY0/BHr5O+JpVqauosdVglVZWp61mddpquvh0YXLsZK7rfh1BpjMvRJV2O+VFxyjOzaYkN4fi3Bz2r1/j8pjDWze3R7H4Bdf9LFKBhq0WnCClzBFCZAgh4qSUCcClKEtSLlGXoZqAlBJL1lGkxYyhSxeXFcg/7s3mgYXbiC1IZVBeIlebM/BNP+I6pXV4bUrrMOcprSoti72igiNjx2ErcN6K2H/KFML+9USHEuzD2zaz8j8vYbcpy9oDLr+KS2fORmha507+tD7dDw5GH3oidlZSU8L3Sd+zJHEJySV1WxTrNDoui7qMqXFTGdJpiEtRs9RUU5KXS3FuDiW5OZTk5TjEoSQ/F1sTlxrDesRy4/Pty0hw/rTxDVqUP7R45Rk5z9bGLT4CDEAycJuUssjlMapYNI6yNWvI/8/rjkCzLiSEwNtuI/DWGY4voJQS85EjbF+2iiM/rqFfQRImm7ne87X3lNZzieoDB8i48y6subl1NwjhCBh7Dh1K5Buvo/V3j2dXS5C8YxsrXnvBcQGNH3MFl99+d6sJRp0+3ReEETDp9D7dUkq2525nSeISVqetxmKve3Hv5tOV6yOuZYTXIKxF5RTnZNcKgiIMFUUtUmbgYODY8Vw6c3aLntMJTU2dDQI+Q7EqP04p8K+HFq/8bwuOq9GoYtEISn/+maz7H6h3m/+0aXieN+icSWk9W7HX1FC2ahXV+w+g8fbGd9xYrPn5ZN53P/YyZc3b0LUrke+9izE6uo1H23hSd+9g+SvPYbUoNy19R17KFbPvRdMKmUmN7dNttVgozc8jM+Mwmw78yqGUXcjiSnwqdHhX6dDbmihmQuAdGIR/pzD8QsPw79QZv05h+HcK41hWFj+/U//MQaPTccvLbxIU0aU5b7epNLcor45F+UOLV7aZRbkqFg0gbTaOXH451qPO0yudURIWRbexo/G+cASeQ4ac9SmtZyM1yclkzL4TS3o6oMwII994Ha9hzlYI2h/p+3bz7cvPYK2pAaDXhSO5cs6D2CwWyouPYfLxxcOrZawvTu7TrTnPi9KIspOWjbIpzsuhrLDApQ1+vei0BHTqTGDniFpBCKsVhM74Boeic7FEuHX5Ujb+bwEnX+t0RiNX3fMwPc8f3py32Rw6fEaKKhYNULVrF6nTG5evfszDlx0hPdkZEkvc1WN45G8XqllLZwHWoiKy7r2Pym3blCd0OsKe/BcBU6e27cCaQObBfSx78Wks1UpKq19oJ8qLjmGzWBAaDT2GDGPkzTPxC21cEZ3dZqOsMJ/inLpxg/LcQobLK9FpDFTbKlmZ8S422bhyAC//ADyDgyj2qOaQPZ2jumOUeVop87RSZbThofPgyugrmRI7hX7B/Zr03SrJy+XgxnVUlhTjH9aZ3hePxuTtVlv1Dn8haDWxEEJ4AL8DRpSsq6VSyqeEENEoZeyBwA7gZimlWQhhBD4HBgOFwDQpZaqr13CHWFRs2kT6zL+73Mc6eCgPB48kwSMEhOCmYVE8e13TPswq7RtpNpP99NOUfLPM8VzgjBmEPvJ/HWY58WjiQb554SnMVfXbqHgFBHLjC6/hE6g4H9dUVtYRguLc7NpZQjal+XlIJ8kagwIvJdZvCADbCn4muex4MZwOv9BOdWYFjllCaBj6kwwq7dLOpqObWJKwhPWZ67FJW53X6BXYiymxU7g65mq89B2iZ0qHvxi0plgIwEtKWS6E0AMbgfuAB4FlUspFQoj3gN1SyneFEHcB/aWUs4UQ04GJUspprl7DHWJhLSjg8KjRYHV+d/TCxXewISgWgGlDuvDvSfFoNB3+s6FyClJKjn3yKXmvvupYRvEeOZLw+fM7TJOn37/6lG3Lv3G63T8sHA9vb0pyc6gqq79Xtys8fHwJC4nmAjkWgcDmaUczyR+/sM54BwY1K1aSW5HLsiPL+CbxG3Ir6yYheOo8uSrmKqbGTqV3UO8mn9uNdPgLgluWoYQQnihicSfwAxAmpbQKIYYD86SUY4UQq2p/3yyE0AE5QIgr7xJ3BbiPPjqXkuX1V99n+oYya/TD2IWGSYMieGXKALSqUJzVlK1ZQ9bD/+cwOjTGxtLl3XcUr6l2ztfP/JOM/fU3A2oMQqPBNzgEv06d8Q8NcwSS/Tp1xi+0kyP20Rp9uq12KxuzNrIkcQkbMjcgqXtp6BfUj6lxUxnbbSye+tPjg5WWSkrNpQR6BDqaOLmRDn9RaFWxEEJoUXxHegBvA68AW6SUPWq3dwF+klL2E0LsA8ZJKTNrtyUBQ6WUBaec8w7gDoCoqKjBaU6a9LQktvIK9s+8A/2eur7+OV6BPD78do56hzC+f2denzYQnbZ1K1JV2gfVBw8q6bY5OQBog4Lo8vZbmAY6835rH3z52APkJruu5zKYTPWKgX9oGD7BIWgb0eGwJr2U/HeU5SdDtB+hs/q3yPiPc7T8KN8c/oZvD39LflXdDERvvTfXdL+GKbFT6BnQk+zybP6z4z+sTluN1W7FS+/FhB4TuGfQPe5cwmqWWGTO3VDHojzyxYvPOHdYCBEHLD7pqRjgSSnl6y6Pc9PMwh/4FngS+PQUsfhRShkvhNgPjD1FLC6QUhY6O6+7Zha/HczljgVbic87wrCcA+jsVvYHxbAxvD8WrY6xfTvx1t/OQ68KxTmFJS+PzLvvoXqPcqcuDAY6v/ACfuOvbuOROeeXD95k72+rnG7v1L0nNz7/WovE2/Le3Y05TVnKCr17YKv06bbYLazPUKxFNh3ddNr2vkF9ySjLoNR8+pJa/5D+fDr2U3fNMpr0B82cu8GlRXnkixe3yIW79oY+C+XG3OWdt1uubrU+6etQKhL9a5eZACKBo7W/ZwJdAGq3+wEtW4HTDOx2pU+2DQ27QmN5r/8E3ho4mbVdzsOiVd7G41f3VoXiHEQfGkrXzxfgc6XSz0GazRx9+GHy//sm7TXLcOAVV7sUgvOvmdRiiRk+F59YlivbkNUi5zwVvUbPZV0v4/3L3+fHiT8ys99MAj0CHdv3F+6vVygA9uTvYWXyylYZ15mQOXeDFviJukIB4AE8j7Kc31JcCiQ1JBTQimIhhAipnVEghDABlwEHgbXA5NrdZnDCindF7WNqt69pitd6a7Enq4T0Y66b8Px6oAnd3lTOKjQeHkTMn0/wXXc5nit45x2yHnwQezs0IAztFsO4ux5Aoz19KWnYpGnEDruoxV7Lo08Q2iAlw6lqbz7W4tb9e3Tx7cIDgx/g18m/8solr3BB2AUNHvNTyk+tOqZmcjXOe1kAPJY5d0NL+fpNB/7XmB1b00iwM7CgdpqjAb6WUq6p6wTvAAAUz0lEQVQUQhwAFgkhngN2Ah/X7v8x8IUQ4gjKjGJ6K46t0ZRVN+w1U1bdLFt5lbMEodEQcu89GKKjyX78caTZTNlPP5OWdZTIt95sd/bmfS4ZQ5e+/dm3bjUluTl4+gfQ95IxBEVGNXxwExAagc9FERQvTwI7lP9x1C19uvVaPeOixzEuehyDPh+E1UWdR7m5vNXH0wwubWB7JBBLI8z/XFHbIe9a4LHG7N9qYiGl3AMMquf5ZOA0yZdSVgOt296rGcR28kGrEdjszic5vTu7tbhHpZ3id8149JERZN59D7bCQqr37CF16jS6vPsOHr3bV1qnT1Aww69v/eZInoM7UfJLGrLKSsXWHHwvjXJrn+5egb3YV7jP6fa4wOb1lW9lWtui/DhXAjuklLkN7ombYhYdmU6+HlzZz3lVa1SgJ2N6ta87R5W2w3PQIKK/XowxVqm7sebkkHrjTZStcW2VfbaiMWjxHqZYvMsaGxVbc9z6+n/r/Ten2zRCw/Re7WIB41R+aWB7KmdgUX4SN9DIJShQxaJRPD8hnkFRp7uNdvI18uEtQ9R0WZU66CMi6PrVV46+GbKyksw5d1P48SftNvDdmniPCAetEjQv/+Mo0tYSN8WNY3zMeGb0mXHa81qhZd7wefQK7OW2sTSBVSid7JzxTOSLF9tcbG+Q2tq3y4FlDe3rOKYjf3jdaVFutdn59WAevx3MxWKzM6RbIBMGReBtVPtHqdSPtNnIe+VVjn32meM5v+sn0fmppzpUb4yW4NiSRCq3K6sdgdPj8Bzo3tn4wcKDrEhaQWFVIV18uzCp5yQivN1aRNnU1FmnFuWRL16sWpQ3lY7Sg1vl3Kbo66/JeeZZh2WM5/nnE/HfN9AFBLTxyNzHyX269RHehN498FzzTmtuUV4di/LIFy9WLcqbgyoWKh2Fii1byLz3PuylSs6/PiqKLu+9izGm9bOD2gv5n+xztF4Nvj0ej+4dp5FUC9DhlVFdbFdRcQNew4bRbfEiDF27AmBJTyd12nQqNp1edXy2cnKRXnkrFemptB6qWKiouAljdDTdFi/Cc+hQAOxlZaTffgdFixa18cjcg7GHP/owxYup+tAxLHmui11V2heqWKiouBGtvz9RH32I/5TakiKbjZx5T5PzwgtIFzb4ZwNCCLxPnl1sVGcXHQlVLFRU3IzQ6wl75mlC5z4KtUHeos+/IOOuu7CVt8uK4hbDc0AIGl8lE6xiRy62cnMbj0ilsagBbhWVNqRszVqOPvww9uO9MXr2IPLddzFERrbxyFqPk/t0+1wahd/lXdt2QO6hWQHuefPm1bEonzdvXktYlD8A/AOQwF7gtloHDdfHqWKhotK2VCckkHHnnViPZgOgDQwk8q238DzvNLecswJ7pYXsF7cizXY0Xjo6z70Aoe8YrWnPgCaJxbx581xalM+bN69ZF24hRARKI7o+UsoqIcTXKG0iPmvoWHUZSkWljfGIiyN68WJMAwYAYDt2jPQZMyj5/vs2HlnroPHU4zVEsdCxV1ip2KG6Np/MvHnzWtuiXAeYaltBeHKiTYRLVLFQUWkH6EJCiPp8Ab5XK42TpMXC0f97hLzXX0fa3WeP4S68L4pw3GuXb8hCujDqPAdp0KJ83rx5zbKOkFJmAa8C6UA2UCKlbMiLClDFQkWl3aAxGgl/9RWC77nb8Vzhe++T9cCD2Kuq2nBkLY8u0ANTv2AArAVVVB9q8z5n7YnGWpQ3GSFEAHAdEA2EA15CiJsac6wqFioq7QghBCFz5hDx2nyE0QhA2apVpN18C5bcs2u5xtsNnfQ6KK1pUX4ZkCKlzJdSWlCMBEc05kBVLFRU2iG+V11F188XoA1W7r6r9+0jdepUqg+cUb+bdoUxyhdDV18AzCklmDPL2nhE7YbWtChPB4YJITyFYs51KUoH0wZRxUJFpZ1iGjBA6Y0RpzTosebmKr0xfv21jUfWcrijT3cHpEGL8nnz5jXLolxK+SewFNiBkjarAT5ozLGqWKiotGP04eF0XbgQ79GjAZBVVWTecy8FH354VvTGcHef7o7AvHnz7CjW5CtP2VQK3Ddv3rxPz+T8UsqnpJS9pJT9pJQ3SylrGnOcWmehotIBkDYbefNf49gnnzie85swgbBnnkbTwXtjlG8+qvTpRsmS8h9/VjrxNrcor45F+bx581SL8uagioXKuUbx0qVkz3va0RvDNGQwkW++2aF7Y9jNNnJe3Iq90oowaun82AVu7dPtJlSLchUVFffhP3kyUR9/jNbPD4Cqv7aTOnUaNUlJbTyy5qMxaPEa2nZ9ulUahyoWKiodDK+hFyi9Mbp1A8CSkUHq9Bso3/hH2w7sDKjbpzvLrX26VRqHKhYqKh0QQ7duSm+M4cMApTdGxqxZHFu4sI1H1jy0PgZHX25biZmqvQVtPCKVU1HFQkWlg6L18yPqgw/wnzZNecJmI/fZ58h59rkO2RvD55K6abQdOZ56NqKKhYpKB0bo9YTNe4pO/3wMNMrXuWjhQjJm34mtrGMVuek7eWGMVQL1lqxyapJL2nhEKiejioWKSgdHCEHgLbfQ5d130HgpbUsrNm4k9YYbMGdktPHomobap7v9ooqFispZgvfIkXT96iv04eEAmI8kkTp1GpXbt7fxyBqP2qe7/aKKhYrKWYRHXCzdlnyNaaDicG0rKiL91tso/u67Nh5Z4xBC4H2J2qe7PaKKhYrKWYYuKIioBZ/he801gNIbI3vuY+TNf61D9Mbw7K/26W6PqGKhonIWojEaCX/5JULuu9fxXOGHH5J13/2Oft/tFaHTKHUXAFZJ+ebsth2QCqC011NRUTkLEUIQfOedGKKjOfroXGRNDWWrV5OalUXA9GmUrfoFc3o6uuBg/CZMwH/iBEQ78ZnyviCMsjXpSLOdii1H8R0VeS706W7XqN5QKirnAFV795Jx113Y8p0Xu3mNGEHke++2G2PC4hVJlG9S2kP7T+yBd60lSAdF9YZSUVFp/5ji44n++mv0Xbo43adi0yaKvvjSjaNyjdqnu32hioWKyjmCvnNnjD26u9yneOlSN42mYdQ+3e0LVSxUVM4hrA308TanplK0aDGW3Fw3jcg1ap/u9oMqFioq5xC6sDDXO0hJzrx5HBk5ipTrJ5P/1ttUHzjQZj5Nap/u9oMqFioq5xD+k69v9L7V/9/e3QdHdZ13HP8+u9qVtFoBAgECOwo2dfALxibg1AbjSkomtTtN7bQdBnfqacYOTqeTuk4aZ2onM3Wm0zaZNmk746Qz6cQzbmILTx3Hbog9ZBqtMGBSQnFM8JSxm+AXDDjIvGkX62336R+7AiGtdAGzuuzd3+cv6Z778hz+0I977r3nvPIKfY88wr7f/wP+r7OLg1/5CtktWygMTe93D+MnGJRwKCxEaki6s5OZn/xk2bbGlSu4/LkfMf/LX6Zp9WpIJE61jRw6xLHuDby1/l5evfEm9v/5fRx7+geMHKn8c4SGq+ZQp3W6Q6dXZ0VqjBcKnNi4kaMbnmTorTepa53LrDtuZ9a6dcTq60/tl89myW3dRjbTQ7Z3M/njZWaBNaNx+XLSnR00d3WRvPxyzC78W6IRWKe76l+dVViISCAfGeG9n/+c/kyGbE+GoX37yu6XaG+nubOTdGcnqRUfxsbcnbwfEVinW2ERJoWFSDgG9+0jm+kl29PDyV27oMycU7EZM0ivWUO6q5P0mjXEZ8x4X9c8vul1+jPFKddn/s5lNN9y6fs63zRTWEx6YrMPAP8OtAEF4Nvu/i9m9jCwHjhc2vUhd3+udMyDwD1AHrjP3TdNdQ2FhUj4Ro4eJbdlC/2ZDLkXtlDI5SbuVFdHauVKmruKdx3JKT4OnEy+f4iDX90BeSc+M0nbF2/A4lXz2FVhMemJzRYAC9x9l5k1A/8D3AGsBbLu/o/j9r8a6AY+AiwE/gv4kLvnJ7uGwkLk4uJDQ5zcuZP+ngzZnh6GDxwou1/9Fb9BuqOTdFcnjcuWYfGzm/fpyFOvcnJn8RuQ2euWnFq3uwooLM76QmbPAo8AqykfFg8CuPvfl37fBDzs7tsnO6fCQuTi5e4Mvvoa2UwP/ZkMAy/vLrtffPZs0h0dNHd10rRqFbFUatJzDr+T451/2gVA4pI08z57fUUeqFdAVRQ5lWkJCzNbBLwALAU+D3wKOAHsBP7S3Y+a2SPAT939e6VjvgM87+6Tzj+gsBCpHiOHD5PdvJn+ngy5F1/EBya+AmvJJKmbbjz1kDwxf/6EfQ4/uofBV48C0Lr+WhoWz6p47ReAwiLwAmZpYDPwt+7+tJnNB/oAB/6G4lDV3Wb2TWD7uLB4zt2/P+589wL3ArS3t6944403Klq/iFx4hYEBctu3k+3JkO3tZeTw4bL7NVx9NemuLpq7Oqm/6irMjIHXjtL3nT3F9itn0/qpa6az9POlsJjy5GYJYCOwyd2/UaZ9EbDR3ZdqGEqkNnmhwMArr5DNZOjvyTC4d2/Z/era2kh3dpDu6CT3UhMjh4qLOM3//AoS8yYfurpIKCwmPXFxIPEx4Ii73z9m+wJ3P1j6+XPAb7r7OjO7BniC0w+4fwJcoQfcIrVl+O236e/tJduTIbdjBwwPT9gnsfgWGq79YwAal7Uw54+WTneZ50phMemJzW4GtgC/oPjqLMBDwJ3A9RSHoV4HPjMmPL4E3A2MAPe7+/NTXUNhIRJtp78iz5DdvJn8sWPFBovT9PG/I9bYgueHyR94nHRH8VlHcvHiMx56uzu5rds4/uyzjPT1kWxvZ9batTQundbhK4VFmBQWIrXD8/niV+Q9PWQzvRC/gvprihMjDu79IUN7fwiMfkXeQbqzi8bl13Porx/m+DPPTDjfvAceYM49d09X+QqLMCksRGrXe3t/ybuP7QePURjsJ7fpr6Bw5pCVNTSUfetq1Ae7nyC1fHmlS4UIhEXVfP4oIjJW45WLSa8qTvkRq2+m9XP/QPNttxJLp0/tM1VQABzbsKGiNUZJVc3EJSIyVnr1JWRfPAAO+eNzuOTr34CR4eJX5Jlejn73u1MeP/ir8hMiykS6sxCRqlVunW5LJmlatYq2Lz1EvLV1yuPjLVXxQd9FQWEhIlXtzHW695/RNvP235vy2Jmf+ERFaooihYWIVLX69hkkF42u033ijHW653z60yQvu6zscU0338yM226blhqjQGEhIlWveU35dbrrWlr44BOP03LXXcRK62nUtbUx9/6/4NJvfROr02Pbs6VXZ0Wk6nnBeefrOxl5dwBi0PbADdS1NJy5jzs+PIwlEmHMVKtXZ0VEwmYxI31z6e6iANltE9fRMDNiyWS1TGl+0VFYiEgkpFbMJ5YqDivlfnaIwsBIyBVFi8JCRCIhlozTdOMCAHwwT27HoZArihaFhYhERvqmhRAvDjNlt72N5wsBR8jZUliISGTEm5OklhfX5c4fH+K93X0hVxQdCgsRiZTmcR/pVfMbnxcThYWIREpifhMNS1oAGD6QY/BXx0OuKBoUFiISOWOnAMmO+UhPzp/CQkQip37xLBILmgAY2HuE4V+fDLmi6qewEJHIMTPdXVxgCgsRiaTUsrnEZiQByL30Dvn+oZArqm4KCxGJJKuL0bx6YfGXESf704PhFlTlFBYiEllNH1mAJeMAZF/Yz5Hvv0b/lrfJ54YDjpTxND+viERWrLGORFuKoTf78eECJ39WnALkxI9fp2XtElLXTr2SnpymOwsRiaz39vQx9Gb/hO0+XOBI9169JXUOFBYiEln9W6d4C6rgZLdPnMpcylNYiEhkjV1itXx7dpoqqX4KCxGJrFjp4fak7Qn9CTxb+pcSkchqXDr1A+xGPeA+awoLEYms5o4PnFo9b7xEW4rUivnTXFH1UliISGTVzW5g7p9eR/2HWk5vjBup5fNoXb8scJhKTtN3FiISaYl5KebevZT8iSHyuWHqZiaJpRJhl1V1FBYiUhPiM5LES3NFybnTMJSIiARSWIiISCCFhYiIBFJYiIhIIIWFiIgEUliIiEgghYWIiARSWIiISCCFhYiIBFJYiIhIIIWFiIgEMncPu4bzZmaHgTdCuHQr0BfCdcOi/kZbLfU3rL72ufutIVz3gqnqsAiLme1095Vh1zFd1N9oq6X+1lJfLzQNQ4mISCCFhYiIBFJYnJ9vh13ANFN/o62W+ltLfb2g9MxCREQC6c5CREQCKSxERCSQwuI8mFk27Bqmw2g/zWyRme0Ju55KKddPM+sws43hVlYZY/q7z8yWjGv7ZzP7YjiVVYaZ9ZrZb4/bdr+ZfSusmqqRwkKkdm0A1o3+YmYx4A+BJ0OrqDK6GdPPknWl7XKWFBYitWv8H9FbgNfdPYxZESrpKeB3zaweineQwEJga4g1VR2FhUiNcvfdQMHMrittiuT/tt39XWAHMDrdxjrgSderoOdEYSFS27qBdWZWB9wO/EfI9VTK2LuoSIZipSksRGpbN7AW+Biw291/HXI9lfIM8FEz+zDQ6O67wi6o2igsRGqYu/8SeBf4KhH+37a7Z4Fe4FEi3M9KUliISDdwJfCDsAupsG7gOopvgck50nQfIiISSHcWIiISSGEhIiKBFBYiIhJIYSEiIoEUFiIiEkhhIZE22Yy5pZlIV57H+R42sy9cmOpEqofCQkREAikspBbUmdljZrbbzJ4ys9TYRjO708x+YWZ7zOxrY7bfama7zOxlM/vJ+JOa2Xoze97MGqejEyJhqgu7AJFpsAS4x923mdmjwJ+NNpjZQuBrwArgKPBjM7sD2Ab8G3CLu+8zs9ljT2hmnwU+Dtzh7oPT1A+R0CgspBa85e7bSj9/D7hvTNsNQK+7HwYws8cpruuQB15w930A7n5kzDF3AfspBsVwpYsXuRhoGEpqwfg5bcb+bpMcY2WOG7UHWARc+v7KEqkeCgupBe1mdlPp5zs5c4W0/wZ+y8xazSxeat8MbC9tvwxg3DDUS8BngP8sDWOJRJ7CQmrB/wJ/Yma7gdnAv442uPtB4EEgA7wM7HL3Z0vDUvcCT5vZy4xbl9rdtwJfAH5kZq3T0w2R8GjWWRERCaQ7CxERCaSwEBGRQAoLEREJpLAQEZFACgsREQmksBARkUAKCxERCfT/Avi8Rv9QQMEAAAAASUVORK5CYII=\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.catplot(x='block',y='yield',hue='variety', data=oatvar, kind='point')\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The obvious way to fit the model fails, apparently because `patsy` doesn't like the response to be called `yield`."
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"try:\n",
" lmod = smf.ols('yield ~ variety + block', oatvar).fit()\n",
" lmod.summary()\n",
"except:\n",
" pass"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Look at the design matrix:"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([[1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
" [1., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
" [1., 1., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
" [1., 1., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0.],\n",
" [1., 1., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0.],\n",
" [1., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
" [1., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
" [1., 0., 1., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
" [1., 0., 1., 0., 0., 0., 0., 0., 0., 1., 0., 0.],\n",
" [1., 0., 1., 0., 0., 0., 0., 0., 0., 0., 1., 0.],\n",
" [1., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
" [1., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
" [1., 0., 0., 1., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
" [1., 0., 0., 1., 0., 0., 0., 0., 0., 1., 0., 0.],\n",
" [1., 0., 0., 1., 0., 0., 0., 0., 0., 0., 1., 0.],\n",
" [1., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
" [1., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],\n",
" [1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0.],\n",
" [1., 0., 0., 0., 1., 0., 0., 0., 0., 1., 0., 0.],\n",
" [1., 0., 0., 0., 1., 0., 0., 0., 0., 0., 1., 0.],\n",
" [1., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 1.],\n",
" [1., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 1., 0., 0., 1., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0.],\n",
" [1., 0., 0., 0., 0., 1., 0., 0., 0., 0., 1., 0.],\n",
" [1., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 1.],\n",
" [1., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 1., 0., 1., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 1., 0., 0., 1., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 1., 0., 0., 0., 1., 0.],\n",
" [1., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 1.],\n",
" [1., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 1., 1., 0., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 1., 0., 1., 0., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 1., 0., 0., 1., 0.],\n",
" [1., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 1.]])"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import patsy\n",
"dm = patsy.dmatrix('~ variety + block', oatvar)\n",
"np.asarray(dm)"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | yield | R-squared: | 0.748 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.649 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 7.542 | \n",
"
\n",
"\n",
" Date: | Tue, 25 Sep 2018 | Prob (F-statistic): | 7.72e-06 | \n",
"
\n",
"\n",
" Time: | 15:44:55 | Log-Likelihood: | -193.59 | \n",
"
\n",
"\n",
" No. Observations: | 40 | AIC: | 411.2 | \n",
"
\n",
"\n",
" Df Residuals: | 28 | BIC: | 431.4 | \n",
"
\n",
"\n",
" Df Model: | 11 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" const | 363.2750 | 20.027 | 18.139 | 0.000 | 322.252 | 404.298 | \n",
"
\n",
"\n",
" x1 | 42.2000 | 23.125 | 1.825 | 0.079 | -5.169 | 89.569 | \n",
"
\n",
"\n",
" x2 | 28.2000 | 23.125 | 1.219 | 0.233 | -19.169 | 75.569 | \n",
"
\n",
"\n",
" x3 | -47.6000 | 23.125 | -2.058 | 0.049 | -94.969 | -0.231 | \n",
"
\n",
"\n",
" x4 | 105.0000 | 23.125 | 4.541 | 0.000 | 57.631 | 152.369 | \n",
"
\n",
"\n",
" x5 | -3.8000 | 23.125 | -0.164 | 0.871 | -51.169 | 43.569 | \n",
"
\n",
"\n",
" x6 | -16.0000 | 23.125 | -0.692 | 0.495 | -63.369 | 31.369 | \n",
"
\n",
"\n",
" x7 | 49.8000 | 23.125 | 2.154 | 0.040 | 2.431 | 97.169 | \n",
"
\n",
"\n",
" x8 | -25.5000 | 18.282 | -1.395 | 0.174 | -62.949 | 11.949 | \n",
"
\n",
"\n",
" x9 | 0.1250 | 18.282 | 0.007 | 0.995 | -37.324 | 37.574 | \n",
"
\n",
"\n",
" x10 | -42.0000 | 18.282 | -2.297 | 0.029 | -79.449 | -4.551 | \n",
"
\n",
"\n",
" x11 | -77.0000 | 18.282 | -4.212 | 0.000 | -114.449 | -39.551 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 1.932 | Durbin-Watson: | 2.562 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.381 | Jarque-Bera (JB): | 1.397 | \n",
"
\n",
"\n",
" Skew: | 0.458 | Prob(JB): | 0.497 | \n",
"
\n",
"\n",
" Kurtosis: | 3.015 | Cond. No. | 9.57 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: yield R-squared: 0.748\n",
"Model: OLS Adj. R-squared: 0.649\n",
"Method: Least Squares F-statistic: 7.542\n",
"Date: Tue, 25 Sep 2018 Prob (F-statistic): 7.72e-06\n",
"Time: 15:44:55 Log-Likelihood: -193.59\n",
"No. Observations: 40 AIC: 411.2\n",
"Df Residuals: 28 BIC: 431.4\n",
"Df Model: 11 \n",
"Covariance Type: nonrobust \n",
"==============================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"const 363.2750 20.027 18.139 0.000 322.252 404.298\n",
"x1 42.2000 23.125 1.825 0.079 -5.169 89.569\n",
"x2 28.2000 23.125 1.219 0.233 -19.169 75.569\n",
"x3 -47.6000 23.125 -2.058 0.049 -94.969 -0.231\n",
"x4 105.0000 23.125 4.541 0.000 57.631 152.369\n",
"x5 -3.8000 23.125 -0.164 0.871 -51.169 43.569\n",
"x6 -16.0000 23.125 -0.692 0.495 -63.369 31.369\n",
"x7 49.8000 23.125 2.154 0.040 2.431 97.169\n",
"x8 -25.5000 18.282 -1.395 0.174 -62.949 11.949\n",
"x9 0.1250 18.282 0.007 0.995 -37.324 37.574\n",
"x10 -42.0000 18.282 -2.297 0.029 -79.449 -4.551\n",
"x11 -77.0000 18.282 -4.212 0.000 -114.449 -39.551\n",
"==============================================================================\n",
"Omnibus: 1.932 Durbin-Watson: 2.562\n",
"Prob(Omnibus): 0.381 Jarque-Bera (JB): 1.397\n",
"Skew: 0.458 Prob(JB): 0.497\n",
"Kurtosis: 3.015 Cond. No. 9.57\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod = sm.OLS(oatvar['yield'], dm).fit()\n",
"lmod.summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Another solution is to create a response with a new name:"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" df | \n",
" sum_sq | \n",
" mean_sq | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" variety | \n",
" 7.0 | \n",
" 77523.575 | \n",
" 11074.796 | \n",
" 8.284 | \n",
" 0.000 | \n",
"
\n",
" \n",
" block | \n",
" 4.0 | \n",
" 33395.500 | \n",
" 8348.875 | \n",
" 6.245 | \n",
" 0.001 | \n",
"
\n",
" \n",
" Residual | \n",
" 28.0 | \n",
" 37433.300 | \n",
" 1336.904 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" df sum_sq mean_sq F PR(>F)\n",
"variety 7.0 77523.575 11074.796 8.284 0.000\n",
"block 4.0 33395.500 8348.875 6.245 0.001\n",
"Residual 28.0 37433.300 1336.904 NaN NaN"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"oatvar['y'] = oatvar['yield']\n",
"lmod = smf.ols('y ~ variety + block', oatvar).fit()\n",
"sm.stats.anova_lm(lmod).round(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Fit with only block"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" df | \n",
" sum_sq | \n",
" mean_sq | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" block | \n",
" 4.0 | \n",
" 33395.500 | \n",
" 8348.875 | \n",
" 2.542 | \n",
" 0.057 | \n",
"
\n",
" \n",
" Residual | \n",
" 35.0 | \n",
" 114956.875 | \n",
" 3284.482 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" df sum_sq mean_sq F PR(>F)\n",
"block 4.0 33395.500 8348.875 2.542 0.057\n",
"Residual 35.0 114956.875 3284.482 NaN NaN"
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod = smf.ols('y ~ block', oatvar).fit()\n",
"sm.stats.anova_lm(lmod).round(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Leave out the first observation:"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" df | \n",
" sum_sq | \n",
" mean_sq | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" variety | \n",
" 7.0 | \n",
" 75901.631 | \n",
" 10843.090 | \n",
" 9.454 | \n",
" 0.0 | \n",
"
\n",
" \n",
" block | \n",
" 4.0 | \n",
" 38017.908 | \n",
" 9504.477 | \n",
" 8.287 | \n",
" 0.0 | \n",
"
\n",
" \n",
" Residual | \n",
" 27.0 | \n",
" 30967.692 | \n",
" 1146.952 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" df sum_sq mean_sq F PR(>F)\n",
"variety 7.0 75901.631 10843.090 9.454 0.0\n",
"block 4.0 38017.908 9504.477 8.287 0.0\n",
"Residual 27.0 30967.692 1146.952 NaN NaN"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"oat1 = oatvar.iloc[1:,]\n",
"lmod = smf.ols('y ~ variety + block', oat1).fit()\n",
"sm.stats.anova_lm(lmod).round(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Change the factor order:"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" df | \n",
" sum_sq | \n",
" mean_sq | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" block | \n",
" 4.0 | \n",
" 38580.641 | \n",
" 9645.160 | \n",
" 8.409 | \n",
" 0.0 | \n",
"
\n",
" \n",
" variety | \n",
" 7.0 | \n",
" 75338.897 | \n",
" 10762.700 | \n",
" 9.384 | \n",
" 0.0 | \n",
"
\n",
" \n",
" Residual | \n",
" 27.0 | \n",
" 30967.692 | \n",
" 1146.952 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" df sum_sq mean_sq F PR(>F)\n",
"block 4.0 38580.641 9645.160 8.409 0.0\n",
"variety 7.0 75338.897 10762.700 9.384 0.0\n",
"Residual 27.0 30967.692 1146.952 NaN NaN"
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod = smf.ols('y ~ block + variety', oat1).fit()\n",
"sm.stats.anova_lm(lmod).round(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Or use type III ANOVA which is like `drop()` in R"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" sum_sq | \n",
" df | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" Intercept | \n",
" 358745.360119 | \n",
" 1.0 | \n",
" 312.781616 | \n",
" 2.244635e-16 | \n",
"
\n",
" \n",
" block | \n",
" 38017.908036 | \n",
" 4.0 | \n",
" 8.286729 | \n",
" 1.699669e-04 | \n",
"
\n",
" \n",
" variety | \n",
" 75338.897321 | \n",
" 7.0 | \n",
" 9.383744 | \n",
" 7.266460e-06 | \n",
"
\n",
" \n",
" Residual | \n",
" 30967.691964 | \n",
" 27.0 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" sum_sq df F PR(>F)\n",
"Intercept 358745.360119 1.0 312.781616 2.244635e-16\n",
"block 38017.908036 4.0 8.286729 1.699669e-04\n",
"variety 75338.897321 7.0 9.383744 7.266460e-06\n",
"Residual 30967.691964 27.0 NaN NaN"
]
},
"execution_count": 15,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"sm.stats.anova_lm(lmod,typ=3)"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"lmod = smf.ols('y ~ variety + block', oatvar).fit()\n",
"sns.residplot(lmod.fittedvalues, lmod.resid)\n",
"plt.xlabel(\"Fitted values\")\n",
"plt.ylabel(\"Residuals\")\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"res = np.sort(lmod.resid)\n",
"n = len(res)\n",
"qq = stats.norm.ppf(np.arange(1,n+1)/(n+1))\n",
"\n",
"fig, ax = plt.subplots()\n",
"ax.scatter(qq, res)\n",
" \n",
"plt.xlabel(\"Normal Quantiles\")\n",
"plt.ylabel(\"Sorted Residuals\")\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Do the interaction test:"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Intercept 363.275\n",
"variety[T.2] 42.200\n",
"variety[T.3] 28.200\n",
"variety[T.4] -47.600\n",
"variety[T.5] 105.000\n",
"variety[T.6] -3.800\n",
"variety[T.7] -16.000\n",
"variety[T.8] 49.800\n",
"block[T.II] -25.500\n",
"block[T.III] 0.125\n",
"block[T.IV] -42.000\n",
"block[T.V] -77.000\n",
"dtype: float64"
]
},
"execution_count": 18,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod.params"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"varcoefs = np.append([0.],lmod.params[1:8])\n",
"varcoefs = np.repeat(varcoefs,5)\n",
"blockcoefs = np.append([0.],lmod.params[8:12])\n",
"blockcoefs = np.tile(blockcoefs,8)"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | y | R-squared: | 0.749 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.638 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 6.718 | \n",
"
\n",
"\n",
" Date: | Tue, 25 Sep 2018 | Prob (F-statistic): | 2.08e-05 | \n",
"
\n",
"\n",
" Time: | 15:44:55 | Log-Likelihood: | -193.47 | \n",
"
\n",
"\n",
" No. Observations: | 40 | AIC: | 412.9 | \n",
"
\n",
"\n",
" Df Residuals: | 27 | BIC: | 434.9 | \n",
"
\n",
"\n",
" Df Model: | 12 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 362.2421 | 20.505 | 17.666 | 0.000 | 320.169 | 404.316 | \n",
"
\n",
"\n",
" variety[T.2] | 44.4099 | 24.146 | 1.839 | 0.077 | -5.134 | 93.954 | \n",
"
\n",
"\n",
" variety[T.3] | 29.6768 | 23.781 | 1.248 | 0.223 | -19.118 | 78.471 | \n",
"
\n",
"\n",
" variety[T.4] | -50.0927 | 24.324 | -2.059 | 0.049 | -100.001 | -0.184 | \n",
"
\n",
"\n",
" variety[T.5] | 110.4985 | 27.335 | 4.042 | 0.000 | 54.412 | 166.585 | \n",
"
\n",
"\n",
" variety[T.6] | -3.9990 | 23.488 | -0.170 | 0.866 | -52.192 | 44.194 | \n",
"
\n",
"\n",
" variety[T.7] | -16.8379 | 23.579 | -0.714 | 0.481 | -65.218 | 31.542 | \n",
"
\n",
"\n",
" variety[T.8] | 52.4079 | 24.402 | 2.148 | 0.041 | 2.339 | 102.477 | \n",
"
\n",
"\n",
" block[T.II] | -24.5878 | 18.709 | -1.314 | 0.200 | -62.975 | 13.800 | \n",
"
\n",
"\n",
" block[T.III] | 0.1205 | 18.564 | 0.006 | 0.995 | -37.970 | 38.211 | \n",
"
\n",
"\n",
" block[T.IV] | -40.4975 | 18.954 | -2.137 | 0.042 | -79.388 | -1.607 | \n",
"
\n",
"\n",
" block[T.V] | -74.2455 | 19.844 | -3.742 | 0.001 | -114.961 | -33.530 | \n",
"
\n",
"\n",
" crossp | 0.0018 | 0.005 | 0.393 | 0.697 | -0.008 | 0.011 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 1.627 | Durbin-Watson: | 2.540 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.443 | Jarque-Bera (JB): | 1.167 | \n",
"
\n",
"\n",
" Skew: | 0.418 | Prob(JB): | 0.558 | \n",
"
\n",
"\n",
" Kurtosis: | 2.986 | Cond. No. | 1.70e+04 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 1.7e+04. This might indicate that there are
strong multicollinearity or other numerical problems."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: y R-squared: 0.749\n",
"Model: OLS Adj. R-squared: 0.638\n",
"Method: Least Squares F-statistic: 6.718\n",
"Date: Tue, 25 Sep 2018 Prob (F-statistic): 2.08e-05\n",
"Time: 15:44:55 Log-Likelihood: -193.47\n",
"No. Observations: 40 AIC: 412.9\n",
"Df Residuals: 27 BIC: 434.9\n",
"Df Model: 12 \n",
"Covariance Type: nonrobust \n",
"================================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"--------------------------------------------------------------------------------\n",
"Intercept 362.2421 20.505 17.666 0.000 320.169 404.316\n",
"variety[T.2] 44.4099 24.146 1.839 0.077 -5.134 93.954\n",
"variety[T.3] 29.6768 23.781 1.248 0.223 -19.118 78.471\n",
"variety[T.4] -50.0927 24.324 -2.059 0.049 -100.001 -0.184\n",
"variety[T.5] 110.4985 27.335 4.042 0.000 54.412 166.585\n",
"variety[T.6] -3.9990 23.488 -0.170 0.866 -52.192 44.194\n",
"variety[T.7] -16.8379 23.579 -0.714 0.481 -65.218 31.542\n",
"variety[T.8] 52.4079 24.402 2.148 0.041 2.339 102.477\n",
"block[T.II] -24.5878 18.709 -1.314 0.200 -62.975 13.800\n",
"block[T.III] 0.1205 18.564 0.006 0.995 -37.970 38.211\n",
"block[T.IV] -40.4975 18.954 -2.137 0.042 -79.388 -1.607\n",
"block[T.V] -74.2455 19.844 -3.742 0.001 -114.961 -33.530\n",
"crossp 0.0018 0.005 0.393 0.697 -0.008 0.011\n",
"==============================================================================\n",
"Omnibus: 1.627 Durbin-Watson: 2.540\n",
"Prob(Omnibus): 0.443 Jarque-Bera (JB): 1.167\n",
"Skew: 0.418 Prob(JB): 0.558\n",
"Kurtosis: 2.986 Cond. No. 1.70e+04\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"[2] The condition number is large, 1.7e+04. This might indicate that there are\n",
"strong multicollinearity or other numerical problems.\n",
"\"\"\""
]
},
"execution_count": 20,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"oatvar['crossp'] = varcoefs * blockcoefs\n",
"tmod = smf.ols(\"y ~ variety + block + crossp\", oatvar).fit()\n",
"tmod.summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Relative efficiency"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"47.04678522492265"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmcrd = smf.ols(\"y ~ variety\", oatvar).fit()\n",
"np.sqrt(lmcrd.scale)"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"36.563691983011935"
]
},
"execution_count": 22,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"np.sqrt(lmod.scale)"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1.6556167903978543"
]
},
"execution_count": 23,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmcrd.scale/lmod.scale"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Latin square"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" run | \n",
" position | \n",
" material | \n",
" wear | \n",
"
\n",
" \n",
" \n",
" \n",
" 1 | \n",
" 1 | \n",
" 1 | \n",
" C | \n",
" 235 | \n",
"
\n",
" \n",
" 2 | \n",
" 1 | \n",
" 2 | \n",
" D | \n",
" 236 | \n",
"
\n",
" \n",
" 3 | \n",
" 1 | \n",
" 3 | \n",
" B | \n",
" 218 | \n",
"
\n",
" \n",
" 4 | \n",
" 1 | \n",
" 4 | \n",
" A | \n",
" 268 | \n",
"
\n",
" \n",
" 5 | \n",
" 2 | \n",
" 1 | \n",
" A | \n",
" 251 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" run position material wear\n",
"1 1 1 C 235\n",
"2 1 2 D 236\n",
"3 1 3 B 218\n",
"4 1 4 A 268\n",
"5 2 1 A 251"
]
},
"execution_count": 24,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"abrasion = pd.read_csv(\"data/abrasion.csv\", index_col=0, \n",
" dtype={'run':'category','position':'category','material':'category','wear':'int'})\n",
"abrasion.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"See the Latin square layout:"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" position | \n",
" 1 | \n",
" 2 | \n",
" 3 | \n",
" 4 | \n",
"
\n",
" \n",
" run | \n",
" | \n",
" | \n",
" | \n",
" | \n",
"
\n",
" \n",
" \n",
" \n",
" 1 | \n",
" 235 | \n",
" 236 | \n",
" 218 | \n",
" 268 | \n",
"
\n",
" \n",
" 2 | \n",
" 251 | \n",
" 241 | \n",
" 227 | \n",
" 229 | \n",
"
\n",
" \n",
" 3 | \n",
" 234 | \n",
" 273 | \n",
" 274 | \n",
" 226 | \n",
"
\n",
" \n",
" 4 | \n",
" 195 | \n",
" 270 | \n",
" 230 | \n",
" 225 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
"position 1 2 3 4\n",
"run \n",
"1 235 236 218 268\n",
"2 251 241 227 229\n",
"3 234 273 274 226\n",
"4 195 270 230 225"
]
},
"execution_count": 25,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"abrasion.pivot(index = 'run', columns='position', values='wear')"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" position | \n",
" 1 | \n",
" 2 | \n",
" 3 | \n",
" 4 | \n",
"
\n",
" \n",
" run | \n",
" | \n",
" | \n",
" | \n",
" | \n",
"
\n",
" \n",
" \n",
" \n",
" 1 | \n",
" C | \n",
" D | \n",
" B | \n",
" A | \n",
"
\n",
" \n",
" 2 | \n",
" A | \n",
" B | \n",
" D | \n",
" C | \n",
"
\n",
" \n",
" 3 | \n",
" D | \n",
" C | \n",
" A | \n",
" B | \n",
"
\n",
" \n",
" 4 | \n",
" B | \n",
" A | \n",
" C | \n",
" D | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
"position 1 2 3 4\n",
"run \n",
"1 C D B A\n",
"2 A B D C\n",
"3 D C A B\n",
"4 B A C D"
]
},
"execution_count": 26,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"abrasion.pivot(index = 'run', columns='position', values='material')"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
""
]
},
"execution_count": 27,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.catplot(x='run',y='wear',hue='position', data=abrasion, kind='point',scale=0.5)"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
""
]
},
"execution_count": 28,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.catplot(x='run',y='wear',hue='material', data=abrasion, kind='point',scale=0.5)"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" sum_sq | \n",
" df | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" Intercept | \n",
" 103836.1 | \n",
" 1.0 | \n",
" 1695.283 | \n",
" 0.000 | \n",
"
\n",
" \n",
" run | \n",
" 986.5 | \n",
" 3.0 | \n",
" 5.369 | \n",
" 0.039 | \n",
"
\n",
" \n",
" position | \n",
" 1468.5 | \n",
" 3.0 | \n",
" 7.992 | \n",
" 0.016 | \n",
"
\n",
" \n",
" material | \n",
" 4621.5 | \n",
" 3.0 | \n",
" 25.151 | \n",
" 0.001 | \n",
"
\n",
" \n",
" Residual | \n",
" 367.5 | \n",
" 6.0 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" sum_sq df F PR(>F)\n",
"Intercept 103836.1 1.0 1695.283 0.000\n",
"run 986.5 3.0 5.369 0.039\n",
"position 1468.5 3.0 7.992 0.016\n",
"material 4621.5 3.0 25.151 0.001\n",
"Residual 367.5 6.0 NaN NaN"
]
},
"execution_count": 29,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod = smf.ols('wear ~ run + position + material', abrasion).fit()\n",
"sm.stats.anova_lm(lmod,typ=3).round(3)"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | wear | R-squared: | 0.951 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.877 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 12.84 | \n",
"
\n",
"\n",
" Date: | Tue, 25 Sep 2018 | Prob (F-statistic): | 0.00283 | \n",
"
\n",
"\n",
" Time: | 15:44:56 | Log-Likelihood: | -47.776 | \n",
"
\n",
"\n",
" No. Observations: | 16 | AIC: | 115.6 | \n",
"
\n",
"\n",
" Df Residuals: | 6 | BIC: | 123.3 | \n",
"
\n",
"\n",
" Df Model: | 9 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 254.7500 | 6.187 | 41.174 | 0.000 | 239.611 | 269.889 | \n",
"
\n",
"\n",
" run[T.2] | -2.2500 | 5.534 | -0.407 | 0.698 | -15.791 | 11.291 | \n",
"
\n",
"\n",
" run[T.3] | 12.5000 | 5.534 | 2.259 | 0.065 | -1.041 | 26.041 | \n",
"
\n",
"\n",
" run[T.4] | -9.2500 | 5.534 | -1.671 | 0.146 | -22.791 | 4.291 | \n",
"
\n",
"\n",
" position[T.2] | 26.2500 | 5.534 | 4.743 | 0.003 | 12.709 | 39.791 | \n",
"
\n",
"\n",
" position[T.3] | 8.5000 | 5.534 | 1.536 | 0.175 | -5.041 | 22.041 | \n",
"
\n",
"\n",
" position[T.4] | 8.2500 | 5.534 | 1.491 | 0.187 | -5.291 | 21.791 | \n",
"
\n",
"\n",
" material[T.B] | -45.7500 | 5.534 | -8.267 | 0.000 | -59.291 | -32.209 | \n",
"
\n",
"\n",
" material[T.C] | -24.0000 | 5.534 | -4.337 | 0.005 | -37.541 | -10.459 | \n",
"
\n",
"\n",
" material[T.D] | -35.2500 | 5.534 | -6.370 | 0.001 | -48.791 | -21.709 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 0.393 | Durbin-Watson: | 2.080 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.822 | Jarque-Bera (JB): | 0.466 | \n",
"
\n",
"\n",
" Skew: | -0.298 | Prob(JB): | 0.792 | \n",
"
\n",
"\n",
" Kurtosis: | 2.414 | Cond. No. | 6.34 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: wear R-squared: 0.951\n",
"Model: OLS Adj. R-squared: 0.877\n",
"Method: Least Squares F-statistic: 12.84\n",
"Date: Tue, 25 Sep 2018 Prob (F-statistic): 0.00283\n",
"Time: 15:44:56 Log-Likelihood: -47.776\n",
"No. Observations: 16 AIC: 115.6\n",
"Df Residuals: 6 BIC: 123.3\n",
"Df Model: 9 \n",
"Covariance Type: nonrobust \n",
"=================================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"---------------------------------------------------------------------------------\n",
"Intercept 254.7500 6.187 41.174 0.000 239.611 269.889\n",
"run[T.2] -2.2500 5.534 -0.407 0.698 -15.791 11.291\n",
"run[T.3] 12.5000 5.534 2.259 0.065 -1.041 26.041\n",
"run[T.4] -9.2500 5.534 -1.671 0.146 -22.791 4.291\n",
"position[T.2] 26.2500 5.534 4.743 0.003 12.709 39.791\n",
"position[T.3] 8.5000 5.534 1.536 0.175 -5.041 22.041\n",
"position[T.4] 8.2500 5.534 1.491 0.187 -5.291 21.791\n",
"material[T.B] -45.7500 5.534 -8.267 0.000 -59.291 -32.209\n",
"material[T.C] -24.0000 5.534 -4.337 0.005 -37.541 -10.459\n",
"material[T.D] -35.2500 5.534 -6.370 0.001 -48.791 -21.709\n",
"==============================================================================\n",
"Omnibus: 0.393 Durbin-Watson: 2.080\n",
"Prob(Omnibus): 0.822 Jarque-Bera (JB): 0.466\n",
"Skew: -0.298 Prob(JB): 0.792\n",
"Kurtosis: 2.414 Cond. No. 6.34\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 30,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%%capture --no-display\n",
"lmod.summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Do the Tukey computation:"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"19.174282907060704"
]
},
"execution_count": 31,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from statsmodels.sandbox.stats.multicomp import get_tukeyQcrit\n",
"get_tukeyQcrit(4,6)*lmod.bse[1]/np.sqrt(2)"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [],
"source": [
"mcoefs = np.append([0],lmod.params[7:10])\n",
"nmats = ['A','B','C','D']"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" Difference | \n",
" lb | \n",
" ub | \n",
"
\n",
" \n",
" \n",
" \n",
" A-B | \n",
" 45.75 | \n",
" 26.58 | \n",
" 64.92 | \n",
"
\n",
" \n",
" B-C | \n",
" -21.75 | \n",
" -40.92 | \n",
" -2.58 | \n",
"
\n",
" \n",
" B-D | \n",
" -10.50 | \n",
" -29.67 | \n",
" 8.67 | \n",
"
\n",
" \n",
" C-D | \n",
" 11.25 | \n",
" -7.92 | \n",
" 30.42 | \n",
"
\n",
" \n",
" A-D | \n",
" 35.25 | \n",
" 16.08 | \n",
" 54.42 | \n",
"
\n",
" \n",
" A-C | \n",
" 24.00 | \n",
" 4.83 | \n",
" 43.17 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Difference lb ub\n",
"A-B 45.75 26.58 64.92\n",
"B-C -21.75 -40.92 -2.58\n",
"B-D -10.50 -29.67 8.67\n",
"C-D 11.25 -7.92 30.42\n",
"A-D 35.25 16.08 54.42\n",
"A-C 24.00 4.83 43.17"
]
},
"execution_count": 33,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import itertools\n",
"dp = set(itertools.combinations(range(0,4),2))\n",
"dcoef = []\n",
"namdiff = []\n",
"for cp in dp:\n",
" dcoef.append(mcoefs[cp[0]] - mcoefs[cp[1]])\n",
" namdiff.append(nmats[cp[0]] + '-' + nmats[cp[1]])\n",
"thsd = pd.DataFrame({'Difference':dcoef},index=namdiff)\n",
"thsd[\"lb\"] = thsd.Difference - get_tukeyQcrit(4,6) * lmod.bse[1]/np.sqrt(2)\n",
"thsd[\"ub\"] = thsd.Difference + get_tukeyQcrit(4,6) * lmod.bse[1]/np.sqrt(2)\n",
"thsd.round(2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Relative efficiency"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"3.8401360544217713"
]
},
"execution_count": 34,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmodr = smf.ols('wear ~ material', abrasion).fit()\n",
"lmodr.scale/lmod.scale"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Balanced incomplete block design"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" treat | \n",
" gain | \n",
" block | \n",
"
\n",
" \n",
" \n",
" \n",
" 1 | \n",
" f | \n",
" 42.2 | \n",
" b1 | \n",
"
\n",
" \n",
" 2 | \n",
" b | \n",
" 32.6 | \n",
" b1 | \n",
"
\n",
" \n",
" 3 | \n",
" c | \n",
" 35.2 | \n",
" b1 | \n",
"
\n",
" \n",
" 4 | \n",
" c | \n",
" 40.9 | \n",
" b2 | \n",
"
\n",
" \n",
" 5 | \n",
" a | \n",
" 40.1 | \n",
" b2 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" treat gain block\n",
"1 f 42.2 b1\n",
"2 b 32.6 b1\n",
"3 c 35.2 b1\n",
"4 c 40.9 b2\n",
"5 a 40.1 b2"
]
},
"execution_count": 35,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"rabbit = pd.read_csv(\"data/rabbit.csv\", index_col=0, \n",
" dtype={'treat':'category','gain':'float','block':'category'})\n",
"rabbit.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Show the arrange of the BIB. I replace the NaNs with spaces for clarity."
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" block | \n",
" b1 | \n",
" b10 | \n",
" b2 | \n",
" b3 | \n",
" b4 | \n",
" b5 | \n",
" b6 | \n",
" b7 | \n",
" b8 | \n",
" b9 | \n",
"
\n",
" \n",
" treat | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" | \n",
"
\n",
" \n",
" \n",
" \n",
" a | \n",
" | \n",
" 37.3 | \n",
" 40.1 | \n",
" | \n",
" 44.9 | \n",
" | \n",
" | \n",
" 45.2 | \n",
" 44 | \n",
" | \n",
"
\n",
" \n",
" b | \n",
" 32.6 | \n",
" | \n",
" 38.1 | \n",
" | \n",
" | \n",
" | \n",
" 37.3 | \n",
" 40.6 | \n",
" | \n",
" 30.6 | \n",
"
\n",
" \n",
" c | \n",
" 35.2 | \n",
" | \n",
" 40.9 | \n",
" 34.6 | \n",
" 43.9 | \n",
" 40.9 | \n",
" | \n",
" | \n",
" | \n",
" | \n",
"
\n",
" \n",
" d | \n",
" | \n",
" 42.3 | \n",
" | \n",
" 37.5 | \n",
" | \n",
" 37.3 | \n",
" | \n",
" 37.9 | \n",
" | \n",
" 27.5 | \n",
"
\n",
" \n",
" e | \n",
" | \n",
" | \n",
" | \n",
" | \n",
" 40.8 | \n",
" 32 | \n",
" 40.5 | \n",
" | \n",
" 38.5 | \n",
" 20.6 | \n",
"
\n",
" \n",
" f | \n",
" 42.2 | \n",
" 41.7 | \n",
" | \n",
" 34.3 | \n",
" | \n",
" | \n",
" 42.8 | \n",
" | \n",
" 51.9 | \n",
" | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
"block b1 b10 b2 b3 b4 b5 b6 b7 b8 b9\n",
"treat \n",
"a 37.3 40.1 44.9 45.2 44 \n",
"b 32.6 38.1 37.3 40.6 30.6\n",
"c 35.2 40.9 34.6 43.9 40.9 \n",
"d 42.3 37.5 37.3 37.9 27.5\n",
"e 40.8 32 40.5 38.5 20.6\n",
"f 42.2 41.7 34.3 42.8 51.9 "
]
},
"execution_count": 36,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"pt = rabbit.pivot(index = 'treat', columns='block', values='gain')\n",
"pt.replace(np.nan,\" \", regex=True)"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.swarmplot(x='block',y='gain',hue='treat',data=rabbit)\n",
"plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.swarmplot(x='treat',y='gain',hue='block',data=rabbit)\n",
"plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" sum_sq | \n",
" df | \n",
" F | \n",
" PR(>F) | \n",
"
\n",
" \n",
" \n",
" \n",
" Intercept | \n",
" 1945.500 | \n",
" 1.0 | \n",
" 193.553 | \n",
" 0.000 | \n",
"
\n",
" \n",
" treat | \n",
" 158.727 | \n",
" 5.0 | \n",
" 3.158 | \n",
" 0.038 | \n",
"
\n",
" \n",
" block | \n",
" 595.735 | \n",
" 9.0 | \n",
" 6.585 | \n",
" 0.001 | \n",
"
\n",
" \n",
" Residual | \n",
" 150.773 | \n",
" 15.0 | \n",
" NaN | \n",
" NaN | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" sum_sq df F PR(>F)\n",
"Intercept 1945.500 1.0 193.553 0.000\n",
"treat 158.727 5.0 3.158 0.038\n",
"block 595.735 9.0 6.585 0.001\n",
"Residual 150.773 15.0 NaN NaN"
]
},
"execution_count": 39,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod = smf.ols('gain ~ treat + block', rabbit).fit()\n",
"sm.stats.anova_lm(lmod,typ=3).round(3)"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"sns.residplot(lmod.fittedvalues, lmod.resid)\n",
"plt.xlabel(\"Fitted values\")\n",
"plt.ylabel(\"Residuals\")\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"res = np.sort(lmod.resid)\n",
"n = len(res)\n",
"qq = stats.norm.ppf(np.arange(1,n+1)/(n+1))\n",
"\n",
"fig, ax = plt.subplots()\n",
"ax.scatter(qq, res)\n",
" \n",
"plt.xlabel(\"Normal Quantiles\")\n",
"plt.ylabel(\"Sorted Residuals\")\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | gain | R-squared: | 0.855 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.720 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 6.318 | \n",
"
\n",
"\n",
" Date: | Tue, 25 Sep 2018 | Prob (F-statistic): | 0.000518 | \n",
"
\n",
"\n",
" Time: | 15:44:57 | Log-Likelihood: | -66.787 | \n",
"
\n",
"\n",
" No. Observations: | 30 | AIC: | 163.6 | \n",
"
\n",
"\n",
" Df Residuals: | 15 | BIC: | 184.6 | \n",
"
\n",
"\n",
" Df Model: | 14 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 36.0139 | 2.589 | 13.912 | 0.000 | 30.496 | 41.531 | \n",
"
\n",
"\n",
" treat[T.b] | -1.7417 | 2.242 | -0.777 | 0.449 | -6.520 | 3.037 | \n",
"
\n",
"\n",
" treat[T.c] | 0.4000 | 2.242 | 0.178 | 0.861 | -4.378 | 5.178 | \n",
"
\n",
"\n",
" treat[T.d] | 0.0667 | 2.242 | 0.030 | 0.977 | -4.712 | 4.845 | \n",
"
\n",
"\n",
" treat[T.e] | -5.2250 | 2.242 | -2.331 | 0.034 | -10.003 | -0.447 | \n",
"
\n",
"\n",
" treat[T.f] | 3.3000 | 2.242 | 1.472 | 0.162 | -1.478 | 8.078 | \n",
"
\n",
"\n",
" block[T.b10] | 3.2972 | 2.796 | 1.179 | 0.257 | -2.662 | 9.257 | \n",
"
\n",
"\n",
" block[T.b2] | 4.1333 | 2.694 | 1.534 | 0.146 | -1.610 | 9.876 | \n",
"
\n",
"\n",
" block[T.b3] | -1.8028 | 2.694 | -0.669 | 0.514 | -7.546 | 3.940 | \n",
"
\n",
"\n",
" block[T.b4] | 8.7944 | 2.796 | 3.145 | 0.007 | 2.835 | 14.754 | \n",
"
\n",
"\n",
" block[T.b5] | 2.3056 | 2.796 | 0.825 | 0.423 | -3.654 | 8.265 | \n",
"
\n",
"\n",
" block[T.b6] | 5.4083 | 2.694 | 2.007 | 0.063 | -0.335 | 11.151 | \n",
"
\n",
"\n",
" block[T.b7] | 5.7778 | 2.796 | 2.066 | 0.057 | -0.182 | 11.737 | \n",
"
\n",
"\n",
" block[T.b8] | 9.4278 | 2.796 | 3.372 | 0.004 | 3.468 | 15.387 | \n",
"
\n",
"\n",
" block[T.b9] | -7.4806 | 2.796 | -2.675 | 0.017 | -13.440 | -1.521 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 2.951 | Durbin-Watson: | 2.921 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.229 | Jarque-Bera (JB): | 1.909 | \n",
"
\n",
"\n",
" Skew: | 0.401 | Prob(JB): | 0.385 | \n",
"
\n",
"\n",
" Kurtosis: | 2.060 | Cond. No. | 12.8 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: gain R-squared: 0.855\n",
"Model: OLS Adj. R-squared: 0.720\n",
"Method: Least Squares F-statistic: 6.318\n",
"Date: Tue, 25 Sep 2018 Prob (F-statistic): 0.000518\n",
"Time: 15:44:57 Log-Likelihood: -66.787\n",
"No. Observations: 30 AIC: 163.6\n",
"Df Residuals: 15 BIC: 184.6\n",
"Df Model: 14 \n",
"Covariance Type: nonrobust \n",
"================================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"--------------------------------------------------------------------------------\n",
"Intercept 36.0139 2.589 13.912 0.000 30.496 41.531\n",
"treat[T.b] -1.7417 2.242 -0.777 0.449 -6.520 3.037\n",
"treat[T.c] 0.4000 2.242 0.178 0.861 -4.378 5.178\n",
"treat[T.d] 0.0667 2.242 0.030 0.977 -4.712 4.845\n",
"treat[T.e] -5.2250 2.242 -2.331 0.034 -10.003 -0.447\n",
"treat[T.f] 3.3000 2.242 1.472 0.162 -1.478 8.078\n",
"block[T.b10] 3.2972 2.796 1.179 0.257 -2.662 9.257\n",
"block[T.b2] 4.1333 2.694 1.534 0.146 -1.610 9.876\n",
"block[T.b3] -1.8028 2.694 -0.669 0.514 -7.546 3.940\n",
"block[T.b4] 8.7944 2.796 3.145 0.007 2.835 14.754\n",
"block[T.b5] 2.3056 2.796 0.825 0.423 -3.654 8.265\n",
"block[T.b6] 5.4083 2.694 2.007 0.063 -0.335 11.151\n",
"block[T.b7] 5.7778 2.796 2.066 0.057 -0.182 11.737\n",
"block[T.b8] 9.4278 2.796 3.372 0.004 3.468 15.387\n",
"block[T.b9] -7.4806 2.796 -2.675 0.017 -13.440 -1.521\n",
"==============================================================================\n",
"Omnibus: 2.951 Durbin-Watson: 2.921\n",
"Prob(Omnibus): 0.229 Jarque-Bera (JB): 1.909\n",
"Skew: 0.401 Prob(JB): 0.385\n",
"Kurtosis: 2.060 Cond. No. 12.8\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 42,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod.summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Tukey pairwise differences:"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [],
"source": [
"mcoefs = np.append([0],lmod.params[1:6])\n",
"nmats = [chr(i) for i in range(ord('a'),ord('f')+1)]"
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" Difference | \n",
" lb | \n",
" ub | \n",
"
\n",
" \n",
" \n",
" \n",
" a-b | \n",
" 1.74 | \n",
" -5.53 | \n",
" 9.02 | \n",
"
\n",
" \n",
" b-c | \n",
" -2.14 | \n",
" -9.42 | \n",
" 5.13 | \n",
"
\n",
" \n",
" b-d | \n",
" -1.81 | \n",
" -9.08 | \n",
" 5.47 | \n",
"
\n",
" \n",
" e-f | \n",
" -8.53 | \n",
" -15.80 | \n",
" -1.25 | \n",
"
\n",
" \n",
" b-e | \n",
" 3.48 | \n",
" -3.79 | \n",
" 10.76 | \n",
"
\n",
" \n",
" b-f | \n",
" -5.04 | \n",
" -12.32 | \n",
" 2.23 | \n",
"
\n",
" \n",
" c-e | \n",
" 5.62 | \n",
" -1.65 | \n",
" 12.90 | \n",
"
\n",
" \n",
" a-f | \n",
" -3.30 | \n",
" -10.58 | \n",
" 3.98 | \n",
"
\n",
" \n",
" c-d | \n",
" 0.33 | \n",
" -6.94 | \n",
" 7.61 | \n",
"
\n",
" \n",
" c-f | \n",
" -2.90 | \n",
" -10.18 | \n",
" 4.38 | \n",
"
\n",
" \n",
" a-e | \n",
" 5.23 | \n",
" -2.05 | \n",
" 12.50 | \n",
"
\n",
" \n",
" a-d | \n",
" -0.07 | \n",
" -7.34 | \n",
" 7.21 | \n",
"
\n",
" \n",
" d-e | \n",
" 5.29 | \n",
" -1.98 | \n",
" 12.57 | \n",
"
\n",
" \n",
" a-c | \n",
" -0.40 | \n",
" -7.68 | \n",
" 6.88 | \n",
"
\n",
" \n",
" d-f | \n",
" -3.23 | \n",
" -10.51 | \n",
" 4.04 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Difference lb ub\n",
"a-b 1.74 -5.53 9.02\n",
"b-c -2.14 -9.42 5.13\n",
"b-d -1.81 -9.08 5.47\n",
"e-f -8.53 -15.80 -1.25\n",
"b-e 3.48 -3.79 10.76\n",
"b-f -5.04 -12.32 2.23\n",
"c-e 5.62 -1.65 12.90\n",
"a-f -3.30 -10.58 3.98\n",
"c-d 0.33 -6.94 7.61\n",
"c-f -2.90 -10.18 4.38\n",
"a-e 5.23 -2.05 12.50\n",
"a-d -0.07 -7.34 7.21\n",
"d-e 5.29 -1.98 12.57\n",
"a-c -0.40 -7.68 6.88\n",
"d-f -3.23 -10.51 4.04"
]
},
"execution_count": 44,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"p = len(mcoefs)\n",
"dp = set(itertools.combinations(range(0,p),2))\n",
"dcoef = []\n",
"namdiff = []\n",
"for cp in dp:\n",
" dcoef.append(mcoefs[cp[0]] - mcoefs[cp[1]])\n",
" namdiff.append(nmats[cp[0]] + '-' + nmats[cp[1]])\n",
"thsd = pd.DataFrame({'Difference':dcoef},index=namdiff)\n",
"thsd[\"lb\"] = thsd.Difference - get_tukeyQcrit(p,lmod.df_resid) * lmod.bse[1]/np.sqrt(2)\n",
"thsd[\"ub\"] = thsd.Difference + get_tukeyQcrit(p,lmod.df_resid) * lmod.bse[1]/np.sqrt(2)\n",
"thsd.round(2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Relative efficiency"
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"3.094507555519525"
]
},
"execution_count": 45,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmodr = smf.ols('gain ~ treat', rabbit).fit()\n",
"lmodr.scale/lmod.scale"
]
},
{
"cell_type": "code",
"execution_count": 46,
"metadata": {},
"outputs": [
{
"data": {
"application/json": {
"Software versions": [
{
"module": "Python",
"version": "3.7.0 64bit [Clang 4.0.1 (tags/RELEASE_401/final)]"
},
{
"module": "IPython",
"version": "6.5.0"
},
{
"module": "OS",
"version": "Darwin 17.7.0 x86_64 i386 64bit"
},
{
"module": "pandas",
"version": "0.23.4"
},
{
"module": "numpy",
"version": "1.15.1"
},
{
"module": "matplotlib",
"version": "2.2.3"
},
{
"module": "seaborn",
"version": "0.9.0"
},
{
"module": "scipy",
"version": "1.1.0"
},
{
"module": "patsy",
"version": "0.5.0"
},
{
"module": "statsmodels",
"version": "0.9.0"
}
]
},
"text/html": [
"Software | Version |
---|
Python | 3.7.0 64bit [Clang 4.0.1 (tags/RELEASE_401/final)] |
IPython | 6.5.0 |
OS | Darwin 17.7.0 x86_64 i386 64bit |
pandas | 0.23.4 |
numpy | 1.15.1 |
matplotlib | 2.2.3 |
seaborn | 0.9.0 |
scipy | 1.1.0 |
patsy | 0.5.0 |
statsmodels | 0.9.0 |
Tue Sep 25 15:44:57 2018 BST |
"
],
"text/latex": [
"\\begin{tabular}{|l|l|}\\hline\n",
"{\\bf Software} & {\\bf Version} \\\\ \\hline\\hline\n",
"Python & 3.7.0 64bit [Clang 4.0.1 (tags/RELEASE\\_401/final)] \\\\ \\hline\n",
"IPython & 6.5.0 \\\\ \\hline\n",
"OS & Darwin 17.7.0 x86\\_64 i386 64bit \\\\ \\hline\n",
"pandas & 0.23.4 \\\\ \\hline\n",
"numpy & 1.15.1 \\\\ \\hline\n",
"matplotlib & 2.2.3 \\\\ \\hline\n",
"seaborn & 0.9.0 \\\\ \\hline\n",
"scipy & 1.1.0 \\\\ \\hline\n",
"patsy & 0.5.0 \\\\ \\hline\n",
"statsmodels & 0.9.0 \\\\ \\hline\n",
"\\hline \\multicolumn{2}{|l|}{Tue Sep 25 15:44:57 2018 BST} \\\\ \\hline\n",
"\\end{tabular}\n"
],
"text/plain": [
"Software versions\n",
"Python 3.7.0 64bit [Clang 4.0.1 (tags/RELEASE_401/final)]\n",
"IPython 6.5.0\n",
"OS Darwin 17.7.0 x86_64 i386 64bit\n",
"pandas 0.23.4\n",
"numpy 1.15.1\n",
"matplotlib 2.2.3\n",
"seaborn 0.9.0\n",
"scipy 1.1.0\n",
"patsy 0.5.0\n",
"statsmodels 0.9.0\n",
"Tue Sep 25 15:44:57 2018 BST"
]
},
"execution_count": 46,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%load_ext version_information\n",
"%version_information pandas, numpy, matplotlib, seaborn, scipy, patsy, statsmodels"
]
}
],
"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": "3.7.0"
}
},
"nbformat": 4,
"nbformat_minor": 2
}