RF-global-type-disorder.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "804762f7-50a0-4d8b-9056-06fc2954f488",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from matminer.featurizers.composition import ElementProperty, Stoichiometry \n",
    "from matminer.featurizers.composition import ValenceOrbital, IonProperty, AtomicOrbitals\n",
    "from matplotlib import pyplot as plt\n",
    "from matminer.datasets import load_dataset\n",
    "from matminer.featurizers.base import MultipleFeaturizer\n",
    "\n",
    "from matminer.featurizers.conversions import DictToObject, StrToComposition\n",
    "from sklearn.ensemble import RandomForestRegressor,RandomForestClassifier\n",
    "from sklearn.model_selection import ShuffleSplit, train_test_split\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.impute import SimpleImputer\n",
    "from scipy import stats\n",
    "from tqdm import tqdm_notebook as tqdm\n",
    "\n",
    "from pymatgen.core.composition import Composition"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69052db3-501c-47b7-bb8a-d7e9ac1d287f",
   "metadata": {},
   "source": [
    "## Loading data and featurizing compositions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d38a132a-cc61-4811-8d5b-a608477c5342",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>formula</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ca0.9 Cr0.2 Cu2 O7 Sr2 Tl0.9</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Cu1 Mo0.3 O6 Sr2 W0.7</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Cl6 K3 Rh1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B0.2 Cr3 P0.8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Cu1 Ge1 Se7 Tb3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        formula  target\n",
       "0  Ca0.9 Cr0.2 Cu2 O7 Sr2 Tl0.9       3\n",
       "1         Cu1 Mo0.3 O6 Sr2 W0.7       1\n",
       "2                    Cl6 K3 Rh1       0\n",
       "3                 B0.2 Cr3 P0.8       1\n",
       "4               Cu1 Ge1 Se7 Tb3       0"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train=pd.read_csv('data/crabnet_data_type/train.csv',usecols=['formula', 'target'])\n",
    "df_test=pd.read_csv('data/crabnet_data_type/test.csv',usecols=['formula', 'target'])\n",
    "df_train.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f2c3fd64-6b22-40f4-8e38-ba2d62492fd8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train['composition']=[Composition(df_train.iloc[i]['formula']).fractional_composition for i in range(len(df_train))]\n",
    "df_test['composition']=[Composition(df_test.iloc[i]['formula']).fractional_composition for i in range(len(df_test))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "31153eb8-3cef-4216-b537-ea1026f5bbcb",
   "metadata": {},
   "outputs": [],
   "source": [
    "featurizer = MultipleFeaturizer([\n",
    "    ElementProperty.from_preset('magpie'),\n",
    "    Stoichiometry(),\n",
    "    ValenceOrbital()\n",
    "])\n",
    "flen=len(featurizer.featurize(df_train['composition'][0]))\n",
    "features_train=np.zeros((len(df_train),flen))\n",
    "features_test=np.zeros((len(df_test),flen))\n",
    "for i,comp in enumerate(df_train['composition']):\n",
    "    features_train[i,:]=featurizer.featurize(comp)\n",
    "for i,comp in enumerate(df_test['composition']):\n",
    "    features_test[i,:]=featurizer.featurize(comp)\n",
    "features_train=np.nan_to_num(features_train, copy=True, nan=0.0, posinf=None, neginf=None)\n",
    "features_test=np.nan_to_num(features_test, copy=True, nan=0.0, posinf=None, neginf=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2ab0e96a-8008-48a1-a55b-88c975d94933",
   "metadata": {},
   "outputs": [],
   "source": [
    "ytrain=df_train['target'].values\n",
    "ytest=df_test['target'].values\n",
    "Xtrain=features_train\n",
    "Xtest=features_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "c6c431e6-faf4-4a13-adcd-bf06dbf33b7d",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import f1_score,balanced_accuracy_score, accuracy_score, matthews_corrcoef\n",
    "from sklearn.metrics import average_precision_score, recall_score, confusion_matrix, ConfusionMatrixDisplay\n",
    "\n",
    "balanced_accuracy=[]\n",
    "f1=[]\n",
    "precision=[]\n",
    "recall=[]\n",
    "roc_auc=[]\n",
    "mc=[]\n",
    "test_size=0.2\n",
    "y_pred_array=np.zeros((5,len(df_test)))\n",
    "\n",
    "for i in range(42,42+5):\n",
    "    RF=RandomForestClassifier()\n",
    "    RF.fit(Xtrain,ytrain)\n",
    "    ypred=RF.predict(Xtest)\n",
    "    y_pred_array[i-42,:]=ypred\n",
    "    f1.append(f1_score(ytest,ypred,average='macro'))\n",
    "    precision.append(average_precision_score(ytest,ypred.reshape(-1, 1),average='macro'))\n",
    "    recall.append(recall_score(ytest,ypred,average='macro'))\n",
    "    balanced_accuracy.append(balanced_accuracy_score(ytest,ypred))\n",
    "    mc.append(matthews_corrcoef(ytest,ypred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "a8499f83",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgcAAAGwCAYAAAA9hsZrAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAABk80lEQVR4nO3dd3hUZdrH8e+k9yGFJARCaKF3UAiuglJdEFx2RQVRlKYoiA1FVsUCCK8CriiyrEIUEMsKoisRUGBFOhi61BBCCQmk10ky8/6RZXBMiAmpZH6f6zrX7pzznDP3jGHOfe7nec4xWCwWCyIiIiL/41DdAYiIiEjNouRAREREbCg5EBERERtKDkRERMSGkgMRERGxoeRAREREbCg5EBERERtO1R1ARTGbzZw/fx5vb28MBkN1hyMiImVksVhIT08nJCQEB4fKu3bNycnBZDKV+zguLi64ublVQEQ1T61JDs6fP09oaGh1hyEiIuUUFxdHgwYNKuXYOTk5NA7zIj6hoNzHCg4OJiYmplYmCLUmOfD29gYgdm8jfLzUW1IV7rm1d3WHYHfM6enVHYJdMWfnVHcIdiXfkscW/mP9Pa8MJpOJ+IQCYvc0wsf7+s8VaelmwrqcxmQyKTmoya50Jfh4OZTrP7iUnpODS3WHYHfMBn3nVclsKP/VpZSRhSrpGvbyNuDlff3vY6Z2d1/XmuRARESktAosZgrK8WShAou54oKpgZQciIiI3TFjwcz1Zwfl2fdGoPq7iIiI2FDlQERE7I4ZM+XpGCjf3jWfkgMREbE7BRYLBZbr7xooz743AnUriIiIiA1VDkRExO5oQGLJlByIiIjdMWOhQMnBNalbQURERGyociAiInZH3QolU3IgIiJ2R7MVSqZuBREREbGhyoGIiNgd8/+W8uxfmyk5EBERu1NQztkK5dn3RqDkQERE7E6BhXI+lbHiYqmJNOZAREREbCg5EBERu2OugKWszp07xwMPPIC/vz8eHh507NiRPXv2WLdbLBamT59OSEgI7u7u9OrVi0OHDtkcIzc3l4kTJxIQEICnpyeDBw/m7NmzNm2Sk5MZOXIkRqMRo9HIyJEjSUlJKVOsSg5ERMTumDFQUI7FjKFM75ecnMwtt9yCs7Mza9eu5fDhw7z99tvUqVPH2mbOnDnMnTuXBQsWsGvXLoKDg+nbty/p6enWNpMnT2bVqlWsXLmSLVu2kJGRwaBBgygoKLC2GT58ONHR0URFRREVFUV0dDQjR44sU7wacyAiIlLJZs+eTWhoKEuWLLGua9SokfX/WywW5s+fz7Rp0xg6dCgAkZGRBAUFsWLFCsaPH09qaioffvghn3zyCX369AFg2bJlhIaGsmHDBvr378+RI0eIiopi+/btdOvWDYDFixcTERHB0aNHadGiRaniVeVARETsjtlS/gUgLS3NZsnNzS32/dasWUPXrl255557CAwMpFOnTixevNi6PSYmhvj4ePr162dd5+rqSs+ePdm6dSsAe/bsIS8vz6ZNSEgIbdu2tbbZtm0bRqPRmhgAdO/eHaPRaG1TGkoORETE7pSnS+HKAhAaGmrt2zcajcyaNavY9zt16hQLFy4kPDyc77//nkcffZRJkybx8ccfAxAfHw9AUFCQzX5BQUHWbfHx8bi4uODr61tim8DAwCLvHxgYaG1TGupWEBERuU5xcXH4+PhYX7u6uhbbzmw207VrV2bOnAlAp06dOHToEAsXLuTBBx+0tjMYbMcyWCyWIut+7/dtimtfmuP8lioHIiJidyqqcuDj42OzXCs5qFevHq1bt7ZZ16pVK86cOQNAcHAwQJGr+4SEBGs1ITg4GJPJRHJycoltLl68WOT9ExMTi1QlSqLkQERE7I7ZYij3Uha33HILR48etVl37NgxwsLCAGjcuDHBwcGsX7/eut1kMrF582Z69OgBQJcuXXB2drZpc+HCBQ4ePGhtExERQWpqKjt37rS22bFjB6mpqdY2paFuBRERkUr21FNP0aNHD2bOnMmwYcPYuXMn//znP/nnP/8JFHYFTJ48mZkzZxIeHk54eDgzZ87Ew8OD4cOHA2A0Ghk9ejTPPPMM/v7++Pn58eyzz9KuXTvr7IVWrVoxYMAAxo4dy6JFiwAYN24cgwYNKvVMBVByICIidui3XQPXu39Z3HTTTaxatYqpU6fy2muv0bhxY+bPn8+IESOsbaZMmUJ2djYTJkwgOTmZbt26sW7dOry9va1t5s2bh5OTE8OGDSM7O5vevXuzdOlSHB0drW2WL1/OpEmTrLMaBg8ezIIFC8oUr8FiqR0PpU5LS8NoNJJ8rAk+3uotqQoDO/ev7hDsjjkt/Y8bSYUxZ2dXdwh2Jd+SxybLalJTU20G+VWkK+eKHw+G4lWOc0VGupk72sZVaqzVSZUDERGxO5brGDfw+/1rM11ii4iIiA1VDkRExO5U9ZiDG42SAxERsTsFFgcKLNdfPC+oFaP1rk3dCiIiImJDlQMREbE7ZgyYy3F9bKZ2lw6UHIiIiN3RmIOSqVtBREREbKhyICIidqf8AxLVrSAiIlKrFI45uP6ugfLseyNQt4KIiIjYUOWggly64MyHM+qxa6MPpmwH6jfJ5em5ZwhvX3hv9k/eCmbT13VIPO+Ms4uFZu2yefiFC7TsnGU9xvnTLix+LYRDO73IMxnocnsaj79xDt+6+UXez5Rr4MmBzTl12J331x2laVvdA/6Kex4+xaiJJ1i9oiGL32oJwH/2riu27Yfzw/nq48YABDfIYvTkY7TplIyzs5k9WwP4YE5LUpKKfz67PRv26Dlu6XeZBk2yMeU6cHivNx/NCeNcjHux7Se+fpI/35/AojcasXppPev62csP0b5bmk3bzd/68+bk5pUaf23h7lnAQ1Mu0GNAKnX88zl5yJ2FLzfg2D4PAJ6ZF0u/Yck2+xzZ68Hku/T9mnGgQLMVrknJQQVIT3Hk6SHhtO+RzhvLTlEnIJ8Lp13w9CmwtqnfJIfHZ5ylXpiJ3BwHVv2zLlPvb8qSrYep419ATpYDL97flCats5n9xQkAIufU4+WHGvPOt8dx+N3f8IdvhOAfnMepw8X/GNur8NapDBh6llPHvGzWP9C3p83rLrdc4smXD7H1hyAAXN3yeeO9PcQc92bq+K4AjHzsBC/P/4VnHupW6++jXlbtbk7lm2XBHDvghaOjhYeePsOMpYcZP6AjudmONm0j+iTRokMGl+Kdiz3W2pWBfDI/1Po6N0cFzdJ66q04GrXIYc6kMJIuOnPH0CTeXHmCsbe35HK8CwC7fvTm7acbWvfJz9PfMmjMwR+pEf8K4+LiGD16NCEhIbi4uBAWFsaTTz7J5cuXqzu0Uvn8vUACQkw8Oz+Olp2yCA410enWDEIamaxt7hiaQufbMqgXZqJRixzGTT9HVrojMf87uR/a6cnFOBeemX+Gxq1yaNwqh2fmneFYtCfRW2xPdLt+9GbPZm/GvnyuSj9nTefmns9zMw7w7uttyEizPRElX3a1Wbr3TGD/bj/izxVeYbXumEJgSDZzX2lL7AlvYk94M396W1q0TaPDTUnV8XFqtJceac2GrwI5c9yDmF89mfdCM4Lqmwhvm2nTzj8olwnTY5jzTDgF+cX/3ORmO5B8ycW6ZGXomqU0XNzM/OnPKfxrRj0O7vDi/GlXls2tR3ycC4MevPrbmWcykJzobF3SU/T9QmHloLxLbVbtn+7UqVN07dqVY8eO8emnn3LixAk++OADfvjhByIiIkhKqvk/zNvXGWneIYs3xjViWLs2TOjbnO+W+12zfZ7JwHfL/PH0KaBJ62zrOgzg7HI1G3VxNePgYOHQzqvJQXKiE/OfC2XKu7G4utfuzLWsHnvhCLu2BBC907/EdnX8crnpT5dYt7q+dZ2zixksBvJMV/9JmEwOFBRA607JxR1GfsPDu7Dr67cnHoPBwrNvneDLxSGcOe5xzX1vH3KJlTt38cHaaMa8cBp3z4JrtpWrHB0tODqBKdf2Zzw3x4E2N2VYX7ePyOCzfQf58KcjTJ5zBqN/XlWHKjegak8hH3/8cVxcXFi3bh3u7oVX0Q0bNqRTp040bdqUadOmsXDhwiL75ebmkpuba32dlpZWpE1VuXDGhW8/DmDouETum3iRo9EeLHypAc4uFvrec/XEsn29D7MeCyM32wG/oDxmrTyB0b/wh7Bll0zcPMx8OCOEh184Dxj41xv1MJsNJCUU/meyWOCtyQ0ZOPIyzTtkEx/nUh0ft0a6rd8FmrVMZ/LIbn/Ytvdd58nOcmTrj4HWdb/ur0NOtiMPP3mMjxeEAxYefvI4jo7gF2C69sEEsDDuxVgO7vIm9jdJwD3jz2MuMPB1ZPA199y4JoD4OFeSE11o1DyLUc+eoXHLLKaNal0Vgd/QsjMdObzbg+FPxnPmuBspiU70ujuZlp2yOBdTOE5m90Yffvq2DhfPuhDc0MRDz11gzucneeLO5jaJsD0qsBgoKEd3YXn2vRFU619HUlIS33//PRMmTLAmBlcEBwczYsQIPvvsMyzF9O3MmjULo9FoXUJDQ4u0qSoWMzRrm80jUy/QrF02A0de5s7hl/nPxwE27TreksH7648yb81xuvZKZ8b4RqRcKjzx1/Ev4O+LTrNjvQ93h7fnLy3akZXuSLN2WTj8rwv36w8DyEp34N6JF6v6I9ZoAUE5jHvuKG/9vR15Jsc/bN938Dk2ra1n0zYtxYVZz7en262JfLnlB77470Y8vfI5ccQbsy5kSzRhegyNW2Qx+6lw67pmbTIY8tAF3p7SDEqY8hX1WRDRW+sQe9yDzf8JYMYTzen8p1Satsm45j5y1ZxJYRgM8OneQ3wbs4+7H7nExlW+1r/ZzWt82fmDkdij7uxYb+TvDzSlfpNcbu5dfRdTNUXB/wYklmepzaq1cnD8+HEsFgutWrUqdnurVq1ITk4mMTGRwMBAm21Tp07l6aeftr5OS0urtgTBLzCfsOY5NutCw3PY8p3RZp2bh5n6jU3Ub2yiVZcsHr6lFVGf+nHfxAQAuvRKZ+m2I6RedsTRCbyMBdzXoQ3BoYUVkuifvfl1ryeDGnWwOe4TdzbnjqHJPPfOmUr8lDVXs1Zp+PqbeGf5dus6RycLbTsnc9ewOO7u3gezufAE1aZTMqGNs5j9Qocix/llewBjhtyKTx0TBfkGMjOcWbZuE/HnNejzWh57OYbuvZN57v42XIq/Oquj7U3p1PHP4+P/7rGuc3SCMVNPc/eoC4zq1bnY45045EmeyUD9sBxOHvIqto1cdSHWlef+Fo6rewGe3maSEpx5ceFp4uOKn2GTlOBMwjln6jfOLXa7yBXV3q1QkisVA4Oh6JWHq6srrq41Y4pZ65syiTtpG8u5U64E1i+5b89igbzcotnnla6G6C1epFxyonu/wix/wutnGfX81avdy/HOvDi8KS9+cJqWnbKKHMde7Nvpx4R7ImzWTZ5+iLOnPflyaSNrYgDQb8g5jh/2Iea49zWPl5ZS2F3T/qbLGP1M7NgceM229svCY6/E0KNvEs+PaMPFs242W39YHcAvP9smx28sOcyPX9dl3ZfX/j7DwrNxdrGQlKgus7LIzXYkN9sRL2M+XXqm8a8ZIcW28/bNp269PJISip85Yk/MFgfM5ZitYK7lsxWqNTlo1qwZBoOBw4cPc/fddxfZ/uuvv+Lr60tAQEDRnWuQoeMSeGpwcz79RyC33ZXC0V88+G6ZP5P/7ywAOVkOrHgniIh+qfgF5ZGW5MS3kQFcuuDMrXelWI/z/Uo/GobnYPTP58geTxa+XJ+/jEsktFlhlh/YIA+4mnC4eZoBCAkzUTfEfgcZZWc5EXvS9mSfk+1IWqqzzXp3z3z+1Deef81tUexx+gw+R1yMJ6nJLrRqn8K4Z4+yenkY52I9KzX+G9Hjr8bQ665LvPZoC7IzHfH937iMzHRHTLmOpKc4k55iewIqyHcgOdHFei+Eeg1zuH1wIrs2+ZKa7ERYs2zGTI3lxCFPDu+5dvImV3XpmYbBAHEnXanfyMSYl85x9qQb6z7zx82jgJHPxLPluzokXXQiKNTEwy9cIDXZiZ/XGv/44LVcebsGCnSfg8rj7+9P3759ef/993nqqadsxh3Ex8ezfPlyHnzwwWIrBzVJi47ZvPxhDEtm1WP5vGCCQ008+to57hhaOBjRwcHC2ROuvP5FI9KSnPD2LaB5hyzeXnWcRi2udkecPenKkln1SE9xJCjUxP2TLjJ0XGJ1faxap2f/eAA2f1/8ALkGYZmMeuI4XsY8Es6789mHjVm9PKwqQ7xhDBpROO5lzorDNuvfntKUDV+VrtKSl2egY49UhjwUj7tnAYkXXNi50Zfl7zawqfbItXn6FPDwCxcIqJdHeoojP39XhyWz61GQb8DRyUCjljn0+VsMnj4FJCU4sW+rFzMfa0R25h+PzRH7ZrAUN9qvCh0/fpwePXrQqlUr3njjDRo3bsyhQ4d47rnnyM3NZfv27fj5XXta4BVpaWkYjUaSjzXBx7t2DxSpKQZ27l/dIdgdc1p6dYdgV8zZuvNoVcq35LHJsprU1FR8fHwq5T2unCsW7e2Cu9f1Xx9nZ+QzvvOeSo21OlX7WTQ8PJzdu3fTtGlT7r33Xpo2bcq4ceO4/fbb2bZtW6kSAxERkbLQTZBKViMGJIaFhbFkyZLqDkNERESoIcmBiIhIVSr/sxVUORAREalVzBgwl3CDrtLsX5spORAREbujykHJavenExERkTJT5UBEROxO+W+CVLuvrZUciIiI3TFbDJjL8WTF8ux7I6jdqY+IiIiUmSoHIiJid8zl7FbQTZBERERqmfI/lbF2Jwe1+9OJiIhImalyICIidqcAAwXluJFRefa9ESg5EBERu6NuhZLV7k8nIiIiZabKgYiI2J0Cytc1UFBxodRISg5ERMTuqFuhZEoORETE7ujBSyWr3Z9OREREykyVAxERsTsWDJjLMebAoqmMIiIitYu6FUpWuz+diIiIlJkqByIiYnf0yOaSKTkQERG7U1DOpzKWZ98bQe3+dCIiIlJmqhyIiIjdUbdCyVQ5EBERu2PGodxLWUyfPh2DwWCzBAcHW7dbLBamT59OSEgI7u7u9OrVi0OHDtkcIzc3l4kTJxIQEICnpyeDBw/m7NmzNm2Sk5MZOXIkRqMRo9HIyJEjSUlJKfP3o+RARESkCrRp04YLFy5YlwMHDli3zZkzh7lz57JgwQJ27dpFcHAwffv2JT093dpm8uTJrFq1ipUrV7JlyxYyMjIYNGgQBQVXn/QwfPhwoqOjiYqKIioqiujoaEaOHFnmWNWtICIidqfAYqCgHF0DV/ZNS0uzWe/q6oqrq2ux+zg5OdlUC66wWCzMnz+fadOmMXToUAAiIyMJCgpixYoVjB8/ntTUVD788EM++eQT+vTpA8CyZcsIDQ1lw4YN9O/fnyNHjhAVFcX27dvp1q0bAIsXLyYiIoKjR4/SokWLUn8+VQ5ERMTuXBlzUJ4FIDQ01FrCNxqNzJo165rvefz4cUJCQmjcuDH33Xcfp06dAiAmJob4+Hj69etnbevq6krPnj3ZunUrAHv27CEvL8+mTUhICG3btrW22bZtG0aj0ZoYAHTv3h2j0WhtU1qqHIiIiN2xlPOpjJb/7RsXF4ePj491/bWqBt26dePjjz+mefPmXLx4kTfeeIMePXpw6NAh4uPjAQgKCrLZJygoiNjYWADi4+NxcXHB19e3SJsr+8fHxxMYGFjkvQMDA61tSkvJgYiIyHXy8fGxSQ6u5c4777T+/3bt2hEREUHTpk2JjIyke/fuABgMtt0cFoulyLrf+32b4tqX5ji/p24FERGxOwUYyr2Uh6enJ+3ateP48ePWcQi/v7pPSEiwVhOCg4MxmUwkJyeX2ObixYtF3isxMbFIVeKPKDkQERG7Y7aUd9xB+d4/NzeXI0eOUK9ePRo3bkxwcDDr16+3bjeZTGzevJkePXoA0KVLF5ydnW3aXLhwgYMHD1rbREREkJqays6dO61tduzYQWpqqrVNaalbQUREpJI9++yz3HXXXTRs2JCEhATeeOMN0tLSeOihhzAYDEyePJmZM2cSHh5OeHg4M2fOxMPDg+HDhwNgNBoZPXo0zzzzDP7+/vj5+fHss8/Srl076+yFVq1aMWDAAMaOHcuiRYsAGDduHIMGDSrTTAVQciAiInbIXM4BiWXd9+zZs9x///1cunSJunXr0r17d7Zv305YWBgAU6ZMITs7mwkTJpCcnEy3bt1Yt24d3t7e1mPMmzcPJycnhg0bRnZ2Nr1792bp0qU4Ojpa2yxfvpxJkyZZZzUMHjyYBQsWlPnzGSwWSzmLIzVDWloaRqOR5GNN8PFWb0lVGNi5f3WHYHfMael/3EgqjDk7u7pDsCv5ljw2WVaTmppaqkF+1+PKuWLkxvtx8XK57uOYMkx8cvunlRprddJZVERERGyoW0FEROxORd0hsbZSciAiInanqscc3GhqXXJwz+39cHIo/g5VUrHe2v5FdYdgdyY3Ktt0JCkfB0/P6g7BrjhYTJBZ3VEI1MLkQERE5I+Yufp8hOvdvzZTciAiInbHgqFcJ3iLkgMREZHa5bdPVrze/Wuz2j2iQkRERMpMlQMREbE7mq1QMiUHIiJid9StULLanfqIiIhImalyICIidsdcztkKmsooIiJSy6hboWTqVhAREREbqhyIiIjdUeWgZEoORETE7ig5KJm6FURERMSGKgciImJ3VDkomZIDERGxOxbKNx3RUnGh1EhKDkRExO6oclAyjTkQERERG6ociIiI3VHloGRKDkRExO4oOSiZuhVERETEhioHIiJid1Q5KJmSAxERsTsWiwFLOU7w5dn3RqBuBREREbGhyoGIiNgdM4Zy3QSpPPveCJQciIiI3dGYg5KpW0FERERsqHIgIiJ2RwMSS6bkQERE7I66FUqm5EBEROyOKgcl05gDERERsaHKgYiI2B1LObsVanvlQMmBiIjYHQtgsZRv/9pM3QoiIiJiQ5UDERGxO2YMGHSHxGtSciAiInZHsxVKpm4FERERsaHKgYiI2B2zxYBBN0G6JiUHIiJidyyWcs5WqOXTFdStICIiIjZUORAREbujAYklU3IgIiJ2R8lByZQcVILhY48xYuwJm3XJl1144M4+ADz18j76DDpns/3XA3V4ZnQPm3Ut2yXz4GPHaNEmhfx8A6eO+fDK5Jsw5TpW7geo4dbOC+X7d0Jt1nkHmHh9924A0hOdWfNmGEd/qkN2miNNb07jr6/GULdxjrV9WoIza2YVtsnNdCSwSTZ9Hj9Hxz9ftrZZPKYl5w57knHJGQ9jPs3/lMJdL8RiDMqrmg96A7v3iYs88mI8qxYH8MEr9QFw8yhg9LQLRPRPw8c3n4tnXfj6wwC+/TigmqOt+YaNP8st/S7ToEk2plwHDu/14aP/C+NcjLu1zdrjW4vd91+zw/j3vwr/Gzi7mBnz/Gl6DrqEq5uZ6G1G3pvehEvxrlXyOWoSDUgsWY1KDhISEnjppZdYu3YtFy9exNfXlw4dOjB9+nQiIiKqO7wyOX3Si78/0c36uqDAdvvurXWZ/3p76+u8PNs/tJbtknntnV18sbQpH7zVmvw8BxqHp2E2V2rYN4zg5llMWHbI+trBsXB0kMUC/xrXEkdnM2MW/4qrVz6b/hXC+w+04YX1v+DqUfgFLns6nJx0R8b861c8/fLY+3VdIp9oTsCa/TRomwlAePdU+k44i09gHqkXXfh6RhhLHmvB5K8OVv0HvoE075DFnx9I4tQhN5v1j756ng49MpgzsSEX41zo3DOdibPOcvmiM9u+N1ZTtDeGdjen8c3yehzb74Wjk4WHnj7DjCWHGH9nJ3KzCy8Whkd0tdmna89kJs88yc/f+1vXjZ8WQ7c7knnzqeakpzgx5oXTTP/nESbd3QGzuXaf7KRsatSAxL/+9a/s27ePyMhIjh07xpo1a+jVqxdJSUnVHVqZmQsMJF92tS5pKbaZeV6eg832jDQXm+1jJx9hzWeN+OLjppw55c35OE9+/rEe+Xn2XTW4wsHRgk9gnnXx8s8HIDHGjdhfvLnnjVM07JBBUNMc7nnjFLmZDuxdc/UK9fReb259KJ6wjhkENMyl38SzuPvkc/aQp7VNrzEXaNQ5A78GuTTukk6fx84R+4s3BXn6Eb0WN48Cnl8Qy/znGpCeavu32qpLFuu/8GP/Ni8unnVh7XJ/Th12J7x9VjVFe+N4aXRrNnwVyJkTHsT86sm8F5oRVN9EeNsMa5vkSy42S/feyezfbiQ+rjBJ8/DKp9/fElg8qxHRW+tw8rAX//dscxo1z6Jjj5Rq+mTV58pshfIs5TFr1iwMBgOTJ0/+TUwWpk+fTkhICO7u7vTq1YtDhw7Z7Jebm8vEiRMJCAjA09OTwYMHc/bsWZs2ycnJjBw5EqPRiNFoZOTIkaSkpJQpvhqTHKSkpLBlyxZmz57N7bffTlhYGDfffDNTp05l4MCB1R1emYWEZvHxf37gw9UbmfLGLwSH2P4Atut8meVRG/jnl5uY+OIBjL651m1G31xatkshNdmFt/61lWVrN/DmB9tp3eHGS5Iqy6XTbrx8c1de+1NnIp9ozqUzhclXvqnwT9rZ9WqJxcERnJwtnNrlY13XpGsav3zrT2aKE2Yz7F3jT77JgWbdU4t9v8wUJ3avrkujLuk4OtfyOUzl8MTMc+z8wYdffvIusu3QTk+690vFPzgPsNChRwb1m+SyZ3PRtlIyD6/CZDg9pfjibx1/Ezf3Sub7LwOt68LbZuLsYmHvljrWdUkJLsQe86B15/RKjbcmKjzBG8qxXP9779q1i3/+85+0b9/eZv2cOXOYO3cuCxYsYNeuXQQHB9O3b1/S06/+95k8eTKrVq1i5cqVbNmyhYyMDAYNGkTBb8rTw4cPJzo6mqioKKKiooiOjmbkyJFlirHGdCt4eXnh5eXF6tWr6d69O66uJfeB5ebmkpt79YSalpZW2SGW2tGDdXh7envOnfHE18/EvY+c4K0Pt/LYfbeRnurC7q112fJDPRIuuBMUksXIR48x8/0dPPngLeTnORJcvzCRGD72OB++05JTx3zoPfAcM9/byYT7b+V8nOcfRFC7hXVMZ8Tc49RtnEP6JWfWvduAd4a244X10QQ1zca3fg7fzglj2MyTuLib2fSvENISXUhLcLYe46EFx4h8ojnTOt6Mg5MZF3czoxf9SkBYrs17rZkVxpaPgzFlOxLWKZ1xHx2p6o97w+g5JJlm7bKZ+OfwYre//1IIk//vLCv2HiY/D8xmA/OfbcChnV5VHOmNzsK4F09zcJc3sceL/y3oMzSR7ExHmy4F37om8kwGMtJsf/ZTLjvjW1fjaKpKRkYGI0aMYPHixbzxxhvW9RaLhfnz5zNt2jSGDh0KQGRkJEFBQaxYsYLx48eTmprKhx9+yCeffEKfPoVj2JYtW0ZoaCgbNmygf//+HDlyhKioKLZv3063boVd24sXLyYiIoKjR4/SokWLUsVZYyoHTk5OLF26lMjISOrUqcMtt9zCiy++yP79+4ttP2vWLGvJxGg0EhoaWmy76rBnWyBbN9Yj9qQP0bsCmP5UYV9g74GFpZ+fNoSw6+dAYk95s3NLEC8/eRP1G2Zy8y2JADj8r2q99quGbPg2lFPHjCye15qzsZ70vSuuWj5TTdL69hQ63JlESMssWvwplXFLCk/YO/9dF0dnC498cJSEU+682KEbU1p158R2H1r1SsbhN1Xu795uSFaqExOWH+KZNfvpNfo8Sya04PyvHjbvdcf4czz7n3089skhHBwtLH86vNbf/OR61A0x8dhr55kzsSF5ucX/rNw9+hItu2Tx8kONeGJAcxa/FsITs87R6Vb7u2otjwmvxNC4RRazn25+zTb9/prAxjUB5JlK8RNvqP039ClO+aoGV2c6pKWl2Sy/vWgtzuOPP87AgQOtJ/crYmJiiI+Pp1+/ftZ1rq6u9OzZk61bCweb7tmzh7y8PJs2ISEhtG3b1tpm27ZtGI1Ga2IA0L17d4xGo7VNadSY5AAKxxycP3+eNWvW0L9/fzZt2kTnzp1ZunRpkbZTp04lNTXVusTF1dyTZm6OE6dPeBMSWnzfavJlNxIuuBPSsHAgXNLlwqpJXIztFVXcaS/qBucU2d/euXqYqdcyi8T/jdwObZfJlLX7mLV/B6/t3MWjHx8hM9kJ/9DC7+5SrCs/Rdbj/v87QfNbUqnfOosBk8/SsH0GWz4Otjm2l18+gU1yaHFrKg+9e4zDG305vVdXur/XrH02vnXzWRB1jO/O7OO7M/vo0COTIaMv8d2Zfbi6FzDqhXj+OT2EHeuNxBxxZ82SADavqcPfHk2s7vBvGI+9dIruvZN4fmSba84waNM1jdCm2UR9EWSzPjnRBWcXC14++Tbr6/jlkXLJGXtjqYAFIDQ01OZCddasWdd8z5UrV7J3795i28THxwMQFGT73y0oKMi6LT4+HhcXF3x9fUtsExgYyO8FBgZa25RGjelWuMLNzY2+ffvSt29fXn75ZcaMGcMrr7zCqFGjbNq5urr+YddDTeHkXEBoo0wORfsVu93baKJuUA5Jlwo/z8Xz7lxKcKV+WKZNu/oNM9m9tW6lx3ujyc81cPGEO01usu1acvcp7INLjHEj7oAXf37mDACm/43uNvwuNTY4WEq8grqyLb80V2N2JvonL8bdbnsl+8y8OOJOuPH5e3VxdARnF0uR2TbmgsLvXf6IhcdejqFH3ySef6ANF8+6XbNl/3sucuyAJzG/2nY5HD/oSZ7JQKdbUvhpbeHgXN+6JsKaZ/HhnLBKjb42i4uLw8fn6nima52X4uLiePLJJ1m3bh1ubtf+72cw2A54tlgsRdb93u/bFNe+NMf5rRqXHPxe69atWb16dXWHUSajJx1hx0+BJF50p45v4ZgDD898NvynAW7u+YwYe5yfNwaTdMmVoHrZPDThKGkpLmzbdOWq1cBXy5owYtxxYo57W8ccNAjLYOYLnar1s9UEX88Io03vZHzr55J+yZn1CxqQk+HIzX8tvAKN/o8/nn55+NbP5cKvHnz1amPa9Uui5W2Fgw2DmmYT0Cibz19swpAXY/H0zePAOn+ObanD2P+NKYiN9uLMPi8ad03Dw1jA5TOurJ3XkICwbBrb4eCtP5Kd6UjsUXebdTlZDqQnX12/b6snY1+6gCnHgYtnnWkfkUmfvyXzz1dDqiPkG8rj00/R665LvPZYS7IzHfENMAGQme5oc98TD698bh1wmcVvNipyjKwMJ9Z9GcjYqadJT3EiPdWJMc/HcvqYB9Fb61TRJ6k5KuomSD4+PjbJwbXs2bOHhIQEunTpYl1XUFDAf//7XxYsWMDRo0eBwiv/evXqWdskJCRYqwnBwcGYTCaSk5NtqgcJCQn06NHD2ubixYtF3j8xMbFIVaIkNSY5uHz5Mvfccw+PPPII7du3x9vbm927dzNnzhyGDBlS3eGViX9gDlPeiManjonUZBeOHqzD06MjSIx3x8W1gLBm6dzx53N4eueRfMmV/Xv8efPFTmRnXf3P8fXKxri4mBn71BG8ffKIOe7N3yfeTPw5+x6MCJBywZWPJzUnM9kJL788wjpl8NSqA/g1KOzrS01wZvUbjUi/5IxPYB43DU2g38SrU30cnS2MX3KEb2aHsXhMS0yZjgSE5TD87RO0vj0FAGc3M/uj/Fk7LxRTliM+gSZa9kzhwXfP4uSqK93rMeuxMB558QLPL4jFu04BCedcWDq7Ht9+7P/HO9u5QSMKf+znLLed1vb2883Y8NXVEnLPgZfAAJu+Kf7GUotmNKYg38DUd47h4mZm3zYjb49rZZ/3OPht38D17l8GvXv35sCBAzbrHn74YVq2bMnzzz9PkyZNCA4OZv369XTqVHgRaDKZ2Lx5M7NnzwagS5cuODs7s379eoYNGwbAhQsXOHjwIHPmzAEgIiKC1NRUdu7cyc033wzAjh07SE1NtSYQpWGwWGrGUJTc3FymT5/OunXrOHnyJHl5eYSGhnLPPffw4osv4u7uXuL+aWlpGI1G+tR/FCeHG6O74Ub31pYvqjsEuzO5Uen/cUv5OXgqGa9K+RYTP2Z+Smpqaqmuxq/HlXNFk6XTcPC4dnn/j5izcjg1aka5Yu3VqxcdO3Zk/vz5AMyePZtZs2axZMkSwsPDmTlzJps2beLo0aN4exdO+X3sscf49ttvWbp0KX5+fjz77LNcvnyZPXv24OhYWEW68847OX/+PIsWLQJg3LhxhIWF8c0335Q6thpTOXB1dWXWrFklDuYQERGpraZMmUJ2djYTJkwgOTmZbt26sW7dOmtiADBv3jycnJwYNmwY2dnZ9O7dm6VLl1oTA4Dly5czadIk66yGwYMHs2DBgjLFUmMqB+WlykHVU+Wg6qlyULVUOahaVVk5aLyk/JWDmIfLVzmoyWpM5UBERKSq6KmMJdOcLBEREbGhyoGIiNgfi6FwKc/+tZiSAxERsTvlfbJi7Ritd23qVhAREREbqhyIiIj9qeKbIN1olByIiIjd0WyFkpUqOfjHP/5R6gNOmjTpuoMRERGR6leq5GDevHmlOpjBYFByICIiN4Za3jVQHqVKDmJiYio7DhERkSqjboWSXfdsBZPJxNGjR8nPz6/IeERERCqfpQKWWqzMyUFWVhajR4/Gw8ODNm3acObMGaBwrMGbb75Z4QGKiIhI1SpzcjB16lT27dvHpk2bcHO7+tCKPn368Nlnn1VocCIiIpXDUAFL7VXmqYyrV6/ms88+o3v37hgMV7+c1q1bc/LkyQoNTkREpFLoPgclKnPlIDExkcDAwCLrMzMzbZIFERERuTGVOTm46aab+M9//mN9fSUhWLx4MRERERUXmYiISGXRgMQSlblbYdasWQwYMIDDhw+Tn5/PO++8w6FDh9i2bRubN2+ujBhFREQqlp7KWKIyVw569OjBzz//TFZWFk2bNmXdunUEBQWxbds2unTpUhkxioiISBW6rmcrtGvXjsjIyIqORUREpErokc0lu67koKCggFWrVnHkyBEMBgOtWrViyJAhODnpOU4iInID0GyFEpX5bH7w4EGGDBlCfHw8LVq0AODYsWPUrVuXNWvW0K5duwoPUkRERKpOmcccjBkzhjZt2nD27Fn27t3L3r17iYuLo3379owbN64yYhQREalYVwYklmepxcpcOdi3bx+7d+/G19fXus7X15cZM2Zw0003VWhwIiIilcFgKVzKs39tVubKQYsWLbh48WKR9QkJCTRr1qxCghIREalUus9BiUqVHKSlpVmXmTNnMmnSJL788kvOnj3L2bNn+fLLL5k8eTKzZ8+u7HhFRESkkpWqW6FOnTo2t0a2WCwMGzbMus7yvzkdd911FwUFBZUQpoiISAXSTZBKVKrkYOPGjZUdh4iISNXRVMYSlSo56NmzZ2XHISIiIjXEdd+1KCsrizNnzmAymWzWt2/fvtxBiYiIVCpVDkpU5uQgMTGRhx9+mLVr1xa7XWMORESkxlNyUKIyT2WcPHkyycnJbN++HXd3d6KiooiMjCQ8PJw1a9ZURowiIiJShcpcOfjxxx/5+uuvuemmm3BwcCAsLIy+ffvi4+PDrFmzGDhwYGXEKSIiUnE0W6FEZa4cZGZmEhgYCICfnx+JiYlA4ZMa9+7dW7HRiYiIVIIrd0gsz1KbXdcdEo8ePQpAx44dWbRoEefOneODDz6gXr16FR6giIiIVK0ydytMnjyZCxcuAPDKK6/Qv39/li9fjouLC0uXLq3o+ERERCqeBiSWqMzJwYgRI6z/v1OnTpw+fZpff/2Vhg0bEhAQUKHBiYiISNW77vscXOHh4UHnzp0rIhYREZEqYaCcT2WssEhqplIlB08//XSpDzh37tzrDkZERESqX6mSg19++aVUB/vtw5mqi8XNFYuja3WHYRcmN76lukOwO4aubao7BLtiif61ukOwKxZLXhW+maYylkQPXhIREfujAYklKvNURhEREandyj0gUURE5IajykGJlByIiIjdKe9dDnWHRBEREbErqhyIiIj9UbdCia6rcvDJJ59wyy23EBISQmxsLADz58/n66+/rtDgREREKoWlApZarMzJwcKFC3n66af585//TEpKCgUFBQDUqVOH+fPnV3R8IiIiUsXKnBy8++67LF68mGnTpuHo6Ghd37VrVw4cOFChwYmIiFSGqn5k88KFC2nfvj0+Pj74+PgQERHB2rVrrdstFgvTp08nJCQEd3d3evXqxaFDh2yOkZuby8SJEwkICMDT05PBgwdz9uxZmzbJycmMHDkSo9GI0Whk5MiRpKSklPn7KXNyEBMTQ6dOnYqsd3V1JTMzs8wBiIiIVLkrd0gsz1IGDRo04M0332T37t3s3r2bO+64gyFDhlgTgDlz5jB37lwWLFjArl27CA4Opm/fvqSnp1uPMXnyZFatWsXKlSvZsmULGRkZDBo0yFrBBxg+fDjR0dFERUURFRVFdHQ0I0eOLPPXU+YBiY0bNyY6OpqwsDCb9WvXrqV169ZlDkBERKTKVfGAxLvuusvm9YwZM1i4cCHbt2+ndevWzJ8/n2nTpjF06FAAIiMjCQoKYsWKFYwfP57U1FQ+/PBDPvnkE/r06QPAsmXLCA0NZcOGDfTv358jR44QFRXF9u3b6datGwCLFy8mIiKCo0eP0qJFi1LHW+bKwXPPPcfjjz/OZ599hsViYefOncyYMYMXX3yR5557rqyHExERuWGlpaXZLLm5uX+4T0FBAStXriQzM5OIiAhiYmKIj4+nX79+1jaurq707NmTrVu3ArBnzx7y8vJs2oSEhNC2bVtrm23btmE0Gq2JAUD37t0xGo3WNqVV5srBww8/TH5+PlOmTCErK4vhw4dTv3593nnnHe67776yHk5ERKTKVdRNkEJDQ23Wv/LKK0yfPr3YfQ4cOEBERAQ5OTl4eXmxatUqWrdubT1xBwUF2bQPCgqyzgiMj4/HxcUFX1/fIm3i4+OtbQIDA4u8b2BgoLVNaV3XfQ7Gjh3L2LFjuXTpEmazudhgREREaqwK6laIi4vDx8fHutrV9dpPBW7RogXR0dGkpKTw73//m4ceeojNmzdbt//+ycYWi+UPn3b8+zbFtS/NcX6vXHdIDAgIUGIgIiJ268rsgytLScmBi4sLzZo1o2vXrsyaNYsOHTrwzjvvEBwcDFDk6j4hIcFaTQgODsZkMpGcnFxim4sXLxZ538TExCJViT9S5uSgcePGNGnS5JqLiIhIjVfeaYwVcBMki8VCbm4ujRs3Jjg4mPXr11u3mUwmNm/eTI8ePQDo0qULzs7ONm0uXLjAwYMHrW0iIiJITU1l586d1jY7duwgNTXV2qa0ytytMHnyZJvXeXl5/PLLL0RFRWlAooiI3BiqeLbCiy++yJ133kloaCjp6emsXLmSTZs2ERUVhcFgYPLkycycOZPw8HDCw8OZOXMmHh4eDB8+HACj0cjo0aN55pln8Pf3x8/Pj2effZZ27dpZZy+0atWKAQMGMHbsWBYtWgTAuHHjGDRoUJlmKsB1JAdPPvlksevfe+89du/eXdbDiYiI1HoXL15k5MiRXLhwAaPRSPv27YmKiqJv374ATJkyhezsbCZMmEBycjLdunVj3bp1eHt7W48xb948nJycGDZsGNnZ2fTu3ZulS5fa3JBw+fLlTJo0yTqrYfDgwSxYsKDM8RosFkuF3CH61KlTdOzYkbS0tIo4XJmlpaVhNBrp3fRJnByv3ecjFafgREx1h2B3DF3aVHcI9iX61+qOwK7kW/LYmP9vUlNTbQb5VaQr54om02bi6OZ23ccpyMnh1IwXKzXW6lRhT2X88ssv8fPzq6jDiYiIVJqKmspYW5U5OejUqZPNlAiLxUJ8fDyJiYm8//77FRqciIiIVL0yJwd33323zWsHBwfq1q1Lr169aNmyZUXFJSIiItWkTMlBfn4+jRo1on///tZ5mSIiIjecKp6tcKMp030OnJyceOyxx0p172gREZGaqqof2XyjKfNNkLp168Yvv/xSGbGIiIhIDVDmMQcTJkzgmWee4ezZs3Tp0gVPT0+b7e3bt6+w4ERERCpNLb/6L49SJwePPPII8+fP59577wVg0qRJ1m0Gg8H6YIeCgoKKj1JERKQiacxBiUqdHERGRvLmm28SE6Mb34iIiNRmpU4OrtxIMSwsrNKCERERqQq6CVLJyjTmoKzPgxYREamR1K1QojIlB82bN//DBCEpKalcAYmIiEj1KlNy8Oqrr2I0GisrFhERkSqhboWSlSk5uO+++wgMDKysWERERKqGuhVKVOqbIGm8gYiIiH0o82wFERGRG54qByUqdXJgNpsrMw4REZEqozEHJSvz7ZNFRERueKoclKjMD14SERGR2k2VAxERsT+qHJRIyYGIiNgdjTkomZKDSrBk5fcE1csqsv7bVY1Z9G57HhxzmJu6XyS4XiaZmc5E76nLkkVtSLrsXszRLLw2Zxtdu13k9Wnd2LYlpPI/QC3h7lnAQ1Mu0GNAKnX88zl5yJ2FLzfg2D4PAB54+gK9hqRQNySPPJOBEwfcWTK7Hkd/8fyDI9uftm0u8re/HCG8aRL+/tm8OuM2tu0I/U0LCw/cf4A7+53Ay8vE0WP+vPfBTcTG1QHAyyuXkcP306XjBQLqZpGW5sq27aFELm9PVpaL9Sj33XOQm7ueo0mTZPLzHPjb8GFV+0FrsLY3p/O3Ry8S3i4L/6A8Xh3TlG3r6hTbdtKsWP484hIfvNqA1R8GARDUIJfIrQeLbT/jsSb89B/fygpdbkBKDirBk+N74eh4Na0Ma5zGzLk/89Om+ri6FdCseQqfftyCUyeMeHnnMf6J/bwycztPjr+9yLHuvuckmkV6fZ56K45GLXKYMymMpIvO3DE0iTdXnmDs7S25HO/CuVNuvPf3BlyIdcHVzcxfxiYya8VJHr6lNalJ+qfxW26u+cTE1GH9D014aepPRbbfM/QwfxlyhLnvRHD2nA/3DzvIzNd+ZMyEu8jOdsbfLxt/v2wWL+nMmTgjgYGZTHxsJ35+WcyYfZv1OE5OZn76uSFHjgbQv8/JqvyINZ6bh5mYw+6s/9yfl/556prtIvql0KJjJpfinW3WJ5534f4u7W3W3Tk8kXsevciujT6VEnONpm6FElX7gMS77rqLPn36FLtt27ZtGAwG9u7dW8VRlU9aqivJSW7W5eaIeM6f9eRAdABZmc5Me+ZP/LSxAefivDl62I+F/+hAeMsU6gbaVhsaN03lL8NOMH9252r6JDcuFzczf/pzCv+aUY+DO7w4f9qVZXPrER/nwqAHLwOwcbUvv/zkTfwZV2KPufPPV+vj6WOmcevsao6+5tm9tz6Ryzvy87aGxWy18JfBv7Ly87b8vK0hsWfq8Pb8CFxd87n9ttMAxJ6pwxtv3saOXQ24EO/Nvv3BRC7rQLebz+HgcHWa9LJP27NqTStOn65TJZ/rRrJ7k5HIt+rzc9S1r/D9g0xMeP0Mc55sTEGe7Y3rzGYDyYnONkuP/in89xtfcrIcKzv8GudKt0J5ltqs2pOD0aNH8+OPPxIbG1tk20cffUTHjh3p3PnGPTk6OZm5vW8c69aGAcXfZdLTMw+zGTIyrmb6rq75PP/yLhbOb09yklsVRVt7ODpacHQCU67tn3hujgNtbsoo0t7J2cyfR1wmI9WBU4eK696RawkOysDPL4e90fWs6/LyHTlwKIhWrRKvuZ+nRx5ZWc6YzdX+M1QrGAwWnpt/mi8XBRF77I//hpu1y6RZ22yiPguogujkRlPt/yoHDRpEYGAgS5cutVmflZXFZ599xujRo4vdLzc3l7S0NJulJoq49TxeXnlsWFvcFRc4uxTw8LhDbNoQSnbW1eRg7BMHOHLQj+0/a4zB9cjOdOTwbg+GPxmPX1AeDg4W7hiaRMtOWfgF5VvbdeuTyupj+/nm1H7+MjaRqfc3Iy1ZXQpl4eubA0Byim0Sm5zihl+dnGL38fbO5f57D7A2qlmlx2cvhk2Ip6AAvv6odM+/6X/vZWKPu3Fkj1clR1ZDWSpgqcWqPTlwcnLiwQcfZOnSpTa3aP7iiy8wmUyMGDGi2P1mzZqF0Wi0LqGhocW2q279/hzL7p1BxQ42dHQ088LLuzA4WHhvXgfr+m49LtChcyKLFrQvso+U3pxJYRgM8OneQ3wbs4+7H7nExlW+mAuuton+2YsJ/Vrw1JBwdm/yZtoHpzH651Vf0Dcyi21lzICl2N9PD/c8Xnt5I2fijCxbqb/xitCsXSZDHk7g7Wcaca0K5W+5uJq5fUgS36+046qBkoMSVXtyAPDII49w+vRpNm3aZF330UcfMXToUHx9i+9fmzp1KqmpqdYlLi6uiqItvcCgLDp2SeD7b8OKbHN0NDP11Z0E1ctk2jO32FQNOnROpF5IJl98+y3f/LCab35YDcCLr+3gzflFB4NJ8S7EuvLc38IZ3KwdD9zUhkmDmuPkbCE+ztXaJjfbkfOnXfl1ryfznm1IQQEMuD+pGqO+8SQnF1YMfH1tx2rUqZNbpJrg7p7HG9N/JCfbmddm9qSgoEb8BN3w2t6cQZ2AfD7ZdoD/nNrDf07tISjUxNi/nyXy5wNF2t86MBlXdzM//NuvGqKVG0GNqJ+2bNmSHj168NFHH3H77bdz8uRJfvrpJ9atW3fNfVxdXXF1db3m9pqg752xpKa4snN7sM36K4lBSP0MXph8K+lptp/jixXN+f4/jWzWLVz6A4vfa8+On22PJX8sN9uR3GxHvIz5dOmZxr9mXLurxgA4u+g5ImURf9GLpCQ3OnW8wMlThScbJ6cC2rW5yEeRnaztPNzzmPHqj+TlOTD9jZ7k5dnfILjK8sO//fnlJ9sZBzOWHeeHr/xY/3nR6kD/ey+xfYOR1CTnItvshYHS1FhK3r82qxHJARQOTHziiSd47733WLJkCWFhYfTu3bu6w7puBoOFvnfGsiGqIebfXB05OJp58bUdNGueyvQXInB0tODrV9gvm57mQn6+g3WWw+8lXnTnYrzm4JdWl55pGAwQd9KV+o1MjHnpHGdPurHuM39c3QsY/uRFtq0zknTRGR/ffAY9dImAenn89G2d6g69xnFzyyOkXrr1dXBQBk0aJ5Ge7kriJU9WrWnJfX87xPnzPpw778199xwkN9eJjf9tBBRWDGa89gNurgXMmXsbHh55eHgUdt+kprlaByXWDcjE2zuXunWzcHCw0KRxYRXn/AVvcnLs90QG4OZRQEijXOvr4NBcmrTOIj3FicTzLqSn2P6cF+QVzk44e8r2t6ReWA5tu2Xw0kN2Pt5DUxlLVGOSg2HDhvHkk0+yYsUKIiMjGTt2LAbDjZubdeySQGBwNuu/s+1SCKibTcSf4gF476MfbbY9/+SfOBBdt8pirO08fQp4+IULBNTLIz3FkZ+/q8OS2fUoyDfg4GigQdNcXvrnaXz88klPduTYPg+eGRpeqpHe9qZ5syTmzNxgfT1+TOH04vU/NOHtdyL44qvWuLoW8MSjO/HyMvHrsQBefOUOsrMLT+jhTZNo1aJwCumSf66xOfZDY4ZwMaFwUNyDI/bTt/fVOfzvv7MWgCkv9mH/waDK+4A3gObts5jz+THr6/GvnAVg/Rf+/xtrUDr9773M5Xhn9v7XDu9t8Bu6Q2LJDBZLzbnFzpgxY/jqq69ITU0lJiaGhg2LH+FfnLS0NIxGI72bPomTY83ubqgtCk7EVHcIdsfQpU11h2Bfon+t7gjsSr4lj435/yY1NRUfn8pJXq6cK9o8OhNH1+ufJl6Qm8OhD16s1FirU40aDTR69GiSk5Pp06dPmRIDERGRMtFshRLVmG4FgIiICGpQIUNERGoznW6uqUZVDkRERKT61ajKgYiISFXQgMSSKTkQERH7o6mMJVK3goiIiNhQ5UBEROyOuhVKpuRARETsj7oVSqRuBREREbGhyoGIiNgddSuUTMmBiIjYH3UrlEjJgYiI2B8lByXSmAMRERGxocqBiIjYHY05KJmSAxERsT/qViiRuhVEREQq2axZs7jpppvw9vYmMDCQu+++m6NHj9q0sVgsTJ8+nZCQENzd3enVqxeHDh2yaZObm8vEiRMJCAjA09OTwYMHc/bsWZs2ycnJjBw5EqPRiNFoZOTIkaSkpJQpXiUHIiJidwwWS7mXsti8eTOPP/4427dvZ/369eTn59OvXz8yMzOtbebMmcPcuXNZsGABu3btIjg4mL59+5Kenm5tM3nyZFatWsXKlSvZsmULGRkZDBo0iIKCAmub4cOHEx0dTVRUFFFRUURHRzNy5MgyxatuBRERsT9V3K0QFRVl83rJkiUEBgayZ88ebrvtNiwWC/Pnz2fatGkMHToUgMjISIKCglixYgXjx48nNTWVDz/8kE8++YQ+ffoAsGzZMkJDQ9mwYQP9+/fnyJEjREVFsX37drp16wbA4sWLiYiI4OjRo7Ro0aJU8apyICIicp3S0tJsltzc3FLtl5qaCoCfnx8AMTExxMfH069fP2sbV1dXevbsydatWwHYs2cPeXl5Nm1CQkJo27attc22bdswGo3WxACge/fuGI1Ga5vSUHIgIiJ258pshfIsAKGhoda+faPRyKxZs/7wvS0WC08//TR/+tOfaNu2LQDx8fEABAUF2bQNCgqybouPj8fFxQVfX98S2wQGBhZ5z8DAQGub0lC3goiI2J8K6laIi4vDx8fHutrV1fUPd33iiSfYv38/W7ZsKbLNYDDYvo3FUmRdkVB+16a49qU5zm+pciAiInKdfHx8bJY/Sg4mTpzImjVr2LhxIw0aNLCuDw4OBihydZ+QkGCtJgQHB2MymUhOTi6xzcWLF4u8b2JiYpGqREmUHIiIiN2pqG6F0rJYLDzxxBN89dVX/PjjjzRu3Nhme+PGjQkODmb9+vXWdSaTic2bN9OjRw8AunTpgrOzs02bCxcucPDgQWubiIgIUlNT2blzp7XNjh07SE1NtbYpDXUriIiI/ani2QqPP/44K1as4Ouvv8bb29taITAajbi7u2MwGJg8eTIzZ84kPDyc8PBwZs6ciYeHB8OHD7e2HT16NM888wz+/v74+fnx7LPP0q5dO+vshVatWjFgwADGjh3LokWLABg3bhyDBg0q9UwFUHIgIiJ2qKpvn7xw4UIAevXqZbN+yZIljBo1CoApU6aQnZ3NhAkTSE5Oplu3bqxbtw5vb29r+3nz5uHk5MSwYcPIzs6md+/eLF26FEdHR2ub5cuXM2nSJOushsGDB7NgwYIyfj5LGe/kUEOlpaVhNBrp3fRJnBz/eECIlF/BiZjqDsHuGLq0qe4Q7Ev0r9UdgV3Jt+SxMf/fpKam2gzyq0hXzhVd7p2Bo4vbdR+nwJTDns+mVWqs1UmVAxERsT96tkKJlByIiIhdqu1PViwPzVYQERERG6ociIiI/bFYCpfy7F+LKTkQERG7U9WzFW406lYQERERG6ociIiI/dFshRIpORAREbtjMBcu5dm/NlO3goiIiNhQ5UBEROyPuhVKpORARETsjmYrlEzJgYiI2B/d56BEGnMgIiIiNlQ5EBERu6NuhZLVuuTAcuEiFoNLdYdhF5yCAqs7BLuTv/dIdYdgV5Ieurm6Q7ArBaYcWPbvqnkzDUgskboVRERExEatqxyIiIj8EXUrlEzJgYiI2B/NViiRuhVERETEhioHIiJid9StUDIlByIiYn80W6FE6lYQERERG6ociIiI3VG3QsmUHIiIiP0xWwqX8uxfiyk5EBER+6MxByXSmAMRERGxocqBiIjYHQPlHHNQYZHUTEoORETE/ugOiSVSt4KIiIjYUOVARETsjqYylkzJgYiI2B/NViiRuhVERETEhioHIiJidwwWC4ZyDCosz743AiUHIiJif8z/W8qzfy2mbgURERGxocqBiIjYHXUrlEzJgYiI2B/NViiRkgMREbE/ukNiiTTmQERERGyociAiInZHd0gsmZIDERGxP+pWKJG6FURERMSGKgciImJ3DObCpTz712ZKDkRExP6oW6FE6lYQERERG6ociIiI/dFNkEqk5EBEROyObp9cMnUriIiIiA0lByIiYn+uDEgsz1IG//3vf7nrrrsICQnBYDCwevXq34VjYfr06YSEhODu7k6vXr04dOiQTZvc3FwmTpxIQEAAnp6eDB48mLNnz9q0SU5OZuTIkRiNRoxGIyNHjiQlJaXMX4+SAxERsT8WwFyOpYy9CpmZmXTo0IEFCxYUu33OnDnMnTuXBQsWsGvXLoKDg+nbty/p6enWNpMnT2bVqlWsXLmSLVu2kJGRwaBBgygoKLC2GT58ONHR0URFRREVFUV0dDQjR44sW7BozIGIiNihihpzkJaWZrPe1dUVV1fXIu3vvPNO7rzzzmKPZbFYmD9/PtOmTWPo0KEAREZGEhQUxIoVKxg/fjypqal8+OGHfPLJJ/Tp0weAZcuWERoayoYNG+jfvz9HjhwhKiqK7du3061bNwAWL15MREQER48epUWLFqX+fKociIiIXKfQ0FBrCd9oNDJr1qwyHyMmJob4+Hj69etnXefq6krPnj3ZunUrAHv27CEvL8+mTUhICG3btrW22bZtG0aj0ZoYAHTv3h2j0WhtU1qqHIiIiP2xUM6bIBX+T1xcHD4+PtbVxVUN/kh8fDwAQUFBNuuDgoKIjY21tnFxccHX17dImyv7x8fHExgYWOT4gYGB1jalpeRARETsTwXdIdHHx8cmOSgPg8Hwu7ewFFlXNAzbNsW1L81xfk/dCiIiItUoODgYoMjVfUJCgrWaEBwcjMlkIjk5ucQ2Fy9eLHL8xMTEIlWJP6LKQSUY9ug5bul3mQZNsjHlOnB4rzcfzQnjXIw7AI5OZh56Ko6uvZKpF5pLZrojv2w1suT/wkhKcLEeZ+LrJ+l0Syp+gSZyshytxzl7yr26PtoN4Z6HTzFq4glWr2jI4rdaAlDHL5eHJx2nU8RlPL3yOPSLLx/Mbsn5OE/rfk9MO0zHmy/jVzeXnGxHjuyrw5J/NOfsac9rvZVda9stnXsevUh4u2z8g/OYProJ276v85sWFh54+gJ/Hn4Zrzr5/PqLJ+9NCyX22NW/3zlfHKNDRIbNcTd97cusxxtXzYeooUbdtpfb28QQVjeF3DxH9p8JZsH33Ym9VOc3rSyMvWM3f7npCN7uuRyKC2TON7dyKsHP2uKD0V/TpckFm2Ov29+UaZ/1tb5++4G1NK93GV/PbNJzXNl5oj7vft+dS+m1/O/eDJTtYrro/hWkcePGBAcHs379ejp16gSAyWRi8+bNzJ49G4AuXbrg7OzM+vXrGTZsGAAXLlzg4MGDzJkzB4CIiAhSU1PZuXMnN998MwA7duwgNTWVHj16lCkmJQeVoN3NqXyzLJhjB7xwdLTw0NNnmLH0MOMHdCQ32xFXNzNN22Ty6XsNOHXEE29jPuP/fppXFv3Kk39pbz3OiYNebFxTl4TzLnjXyeeBSWeZsfQwD/fqjNlcnr/q2iu8dSoDhp7l1DGv36y18Pe50RTkG3j9qY5kZTrxlwdimfHBHh79aw9ycwr/GZw44sPGtcEkXnDH25jHiPEnef29PYy+61Z938Vw8zBz6rAH6z735+XFMUW2D5twkaFjE3j76TDOnnJj+KR4Zq04weiercnOdLS2+265Px+/FWJ9nZujgmbnxhf4YnsbDp8LxNHBzGN9d/LuqG8Z9s695OQ5A/DgrdEMv2U/r/37ds5cqsMjt+9hwcPf8rd595FlunqRsWpXKxZtuMn6OifP0ea9dp8KYcnmzlxK9yDQJ5MnB2xj9v3rGP3Pv1TNh60mVX2HxIyMDE6cOGF9HRMTQ3R0NH5+fjRs2JDJkyczc+ZMwsPDCQ8PZ+bMmXh4eDB8+HAAjEYjo0eP5plnnsHf3x8/Pz+effZZ2rVrZ5290KpVKwYMGMDYsWNZtGgRAOPGjWPQoEFlmqkA1dytMGrUKAwGA48++miRbRMmTMBgMDBq1KiqD6ycXnqkNRu+CuTMcQ9ifvVk3gvNCKpvIrxtJgBZGU5MG9Wan74L4FyMO79Ge7Pw1cY0b5dJ3Xq51uOs/SyIg7t8SDjnxslDXkTODSUwxERQg9xrvbVdc3PP57kZB3j39TZkpDlb14c0zKJV+1Tem9ma44eNnIv15P1ZrXBzL6DngKtlvKivGnBorx8JF9w5+asPH7/fjMB6OQSGZFfHx6nxdm80Evl/Ify81reYrRbuHp3AyneD+XmtL7FH3XnrqTBc3c3cfneSTcvcbAeSE52tS1a6YzHHsy+TIgfy7S8tOZXgx/H4AF779+3U882gVf3E/7WwcP8tB1iyqTMbDzfhZIIf07+8AzfnfPp3OGFzrByTE5czPKxLZq7tgLlPt3bgYFwQ8Sne7D8TTOR/O9E29CKODgVIxdm9ezedOnWyVgaefvppOnXqxMsvvwzAlClTmDx5MhMmTKBr166cO3eOdevW4e3tbT3GvHnzuPvuuxk2bBi33HILHh4efPPNNzg6Xv03s3z5ctq1a0e/fv3o168f7du355NPPilzvNVeOQgNDWXlypXMmzcPd/fCcmNOTg6ffvopDRs2rOboKoaHdz4A6SnX/ro9vPMxmyHzGj+Mru4F9PtbIhfOuJJ4waXYNvbusReOsGtLANE7/bl3zCnremeXwvqfyXQ1FzabDeTnGWjTMYV1qxsUOZarWz59B58j/qw7l+LdKj/4Wia4oQn/oHz2bL46UCvP5MCB7V607prJd8vrWtff/pdk7hiaREqiM7s2+rBsXj2byoKAl5sJgLSswr/F+r7pBHhnsf1EqLVNXoEje0+H0L5hPKt2tbauH9DxOHd2PE5ShjtbjzVk8Y9dbCoLv+XjnsOADsfZfyaYAnMt/29QxY9s7tWrF5YS9jEYDEyfPp3p06dfs42bmxvvvvsu77777jXb+Pn5sWzZsjLFVpxqTw46d+7MqVOn+OqrrxgxYgQAX331FaGhoTRp0qSao6sIFsa9GMvBXd7EHvcotoWzi5mHnzvDpm8CyMqw/U8ycEQ8o6fE4u5p5swJd6aNak1+nsquv3dbvws0a5nO5JHdimw7e9qTi+fdGPXEcRbMaE1OtiN/eSAWv7omfOvaVmEG3nOGh588jrtHAXExnkyb0IX8fH3fZeVXNw+A5Eu2f8/Jl5wIrG+yvt64yo/4My4kJTrTqEU2j7xwniats5k6PLxK463ZLDz15638cjqYk/8bT+DvnQVAUobt+KOkDHeC61y9o17UvnDOJ/twOd2DJkFJPN5vB+H1LvHEkrts9nui/3aGdT+Iu0s++88E8fTHxd+sp1ap4uTgRlMjfvUefvhhlixZYn390Ucf8cgjj5S4T25uLmlpaTZLTTRhegyNW2Qx+6nif+wcncy88M4xHBzgvVeKDsLa+HUATwxuz3P3t+F8rBtT/3HMeiUshQKCchj33FHe+ns78kxFr3YK8h2Y+VxH6odl8dnmjXy19QfadU1i15YAzAW2Ywk2rq3HpPu7M2VMV86f8WDq7H04u6i8et0stt9v4Wyqq+vWrgjgly0+xB51Z/MaP14f34TOt6XTrG1W1cZZg025awvNgi/z98/6FNn2+9OTAYvNd756d2t2nmzAyQQ/1h9oxguf9qNbs3O0CEm02e+TnzrwwIK/8fhHAzGbDUy/58diji72pNorBwAjR45k6tSpnD59GoPBwM8//8zKlSvZtGnTNfeZNWsWr776atUFeR0eezmG7r2Tee7+NlyKL3pjDEcnMy/+4xjBDXJ5YWTrIlUDKByfkJXhxPlYd36N9uKLPbvo0S+Jzd8GVMVHuCE0a5WGr7+Jd5Zvt65zdLLQtnMydw2L4+7ufThxxIeJ90fg4ZWHk5OFtBQX5kZu5/gRo82xsjKcycpw5nycJ0f31+GzzT/S4/YENn9fr6o/1g0tKbFwzIdv3TySEq6O/6jjn09y4rV/dk4ccCfPZKB+4xxOHCy+0mZPnh20hdtanmbcv4aQkHZ1kO3l9MLvxt8rm8u/mVXg65XD5Yxrz2b69XwAefkONPRP5ej5q107qVnupGa5c+ZyHU4n+vKf55fRLvQiB+KCK+FT1RCqHJSoRiQHAQEBDBw4kMjISCwWCwMHDiQgoOST39SpU3n66aetr9PS0ggNDS1hj6pk4bFXYujRN4nnR7Th4tmifdZXEoOQRjm88EAb0lOcizlOMQyocvA7+3b6MeGeCJt1k6cf4uxpT75c2shmpkFWRuH3HBKaSbPWaXyysNkfHl/fd9nFn3Hh8kUnOt+WxslDhScyJ2cz7bpn8OHMkGvuF9YiB2cXC5cTSvnvoday8NxdW+jVOoZH/zWY88m2N9k5l+zNpXQPujWL49iFwt9KJ8cCOjc6z7vfd7/mUZsGJuPsZOZS+rUTryv3ynF2quUVsxo0lbEmqhHJAcAjjzzCE088AcB77733h+2v9XCLmuDxV2PoddclXnu0BdmZjvgGFPaxZqY7Ysp1xMHRwrQFx2jWJpNXxrbEwcFibZOe6kR+ngPBoTncNvAye38ykprkjH+wiXvGncOU48CuTcWNDrdf2VlOxJ70tlmXk+1IWqqzdf2f+sSTmuxCYrwbjZplMO65X9m+KZBfthf+sAbXz+LWfvH8sj2A1GRn/ANz+dtDMZhyHdm1RVWa4rh5FBDS6OqYjeDQXJq0ziI9xYnE8y6s/jCQ+564yLkYN87FuHL/xHhysx3YuLqw37xeWC53/CWJnT/6kJbkRMPmOYx76SzHD7hzeJfXtd7WLjw/+Cf6tz/Bs8sGkJXrgr9XYTdLRo4LuflOgIFPf27Hwz1/Ie5yHeIuGRnVay85eU58v68w4a3vl8qdHY7z87GGpGS60Tgwmcl3buPX8wHsiy2sCLRucJE2DRLZFxtMWrYr9X3TGN9nF3GXfThwphZXDaj6qYw3mhqTHAwYMACTqfAE2b9//2qOpnwGjSi8Q9WcFYdt1r89pSkbvgokIDiXiD6Fd7l6/9v9Nm2mjGjNgR1GTLkOtO2axt2jLuDlk0/KZWcO7vTh6WFtSU2y96uqsvMNyGXM00ep428i+ZIrP3wbwsrFVwe8mnIdaNMphSHDz+Dlk0fKZRcO7vXl2YdvJjW5Ziah1a15hyz+74vj1tePTj8HwLrP/Xj76UZ8/n4QLm5mnphxBm9jAb9GezJ1RDPrTIR8k4GOf0rn7tEJuHmYuXTBhR0/+LB8Xj27v6/E37oV/nYsGrvGZv2rX/bi218Kb+z18U8dcXXO5/nBP+Htlsuhs4FMXDLIOhMhv8CRm5qe494eB/BwyeNiqhc/H23I4h+7YrYUDjfLzXPi9tanGNd7F+7O+VxK92Db8VCmfdaHvIJaPltBSmSwlDS3opKNGjWKlJQUVq9eDVx99OWV+1Tffffd1KlTh6VLl/7hsdLS0jAajdzhcR9OBk31qwoOPt5/3EgqVH7CpeoOwa4kPXRzdYdgVwpMOexbNo3U1NQKe17B7105V/QJfwonx+tP/PMLctlwfF6lxlqdakzlAKiVX7CIiNRAZgsYynFtbFa3QqX5o4rAlYqCiIiIVJ0aVTkQERGpEprKWCIlByIiYofKmRzU8ptE1Yg7JIqIiEjNocqBiIjYH3UrlEjJgYiI2B+zhXJ1DdTy2QrqVhAREREbqhyIiIj9sZgLl/LsX4spORAREfujMQclUnIgIiL2R2MOSqQxByIiImJDlQMREbE/6lYokZIDERGxPxbKmRxUWCQ1kroVRERExIYqByIiYn/UrVAiJQciImJ/zGagHPcqMNfu+xyoW0FERERsqHIgIiL2R90KJVJyICIi9kfJQYnUrSAiIiI2VDkQERH7o9snl0jJgYiI2B2LxYylHE9WLM++NwIlByIiYn8slvJd/WvMgYiIiNgTVQ5ERMT+WMo55qCWVw6UHIiIiP0xm8FQjnEDtXzMgboVRERExIYqByIiYn/UrVAiJQciImJ3LGYzlnJ0K9T2qYzqVhAREREbqhyIiIj9UbdCiZQciIiI/TFbwKDk4FrUrSAiIiI2VDkQERH7Y7EA5bnPQe2uHCg5EBERu2MxW7CUo1vBouRARESklrGYKV/lQFMZRURExI6ociAiInZH3QolU3IgIiL2R90KJao1ycGVLC7fklfNkdgPB7OpukOwO/r7rloFppzqDsGuXPm+q+KqPJ+8ct0DKZ/a/W/RYKkltZGzZ88SGhpa3WGIiEg5xcXF0aBBg0o5dk5ODo0bNyY+Pr7cxwoODiYmJgY3N7cKiKxmqTXJgdls5vz583h7e2MwGKo7nFJLS0sjNDSUuLg4fHx8qjscu6DvvGrp+656N+p3brFYSE9PJyQkBAeHyhsvn5OTg8lU/sqni4tLrUwMoBZ1Kzg4OFRaplkVfHx8bqh/xLWBvvOqpe+76t2I37nRaKz093Bzc6u1J/WKoqmMIiIiYkPJgYiIiNhQclDNXF1deeWVV3B1da3uUOyGvvOqpe+76uk7l/KqNQMSRUREpGKociAiIiI2lByIiIiIDSUHIiIiYkPJgYiIiNhQclCN4uLiGD16NCEhIbi4uBAWFsaTTz7J5cuXqzu0WikhIYHx48fTsGFDXF1dCQ4Opn///mzbtq26Q6t17rrrLvr06VPstm3btmEwGNi7d28VR1V7jRo1CoPBwKOPPlpk24QJEzAYDIwaNarqA5MblpKDanLq1Cm6du3KsWPH+PTTTzlx4gQffPABP/zwAxERESQlJVV3iLXOX//6V/bt20dkZCTHjh1jzZo19OrVS991JRg9ejQ//vgjsbGxRbZ99NFHdOzYkc6dO1dDZLVXaGgoK1euJDs727ouJyeHTz/9lIYNG1ZjZHIjqjW3T77RPP7447i4uLBu3Trc3d0BaNiwIZ06daJp06ZMmzaNhQsXVnOUtUdKSgpbtmxh06ZN9OzZE4CwsDBuvvnmao6sdho0aBCBgYEsXbqUV155xbo+KyuLzz77jJkzZ1ZjdLVT586dOXXqFF999RUjRowA4KuvviI0NJQmTZpUc3Ryo1HloBokJSXx/fffM2HCBGticEVwcDAjRozgs88+q5LHltoLLy8vvLy8WL16Nbm5udUdTq3n5OTEgw8+yNKlS23+jr/44gtMJpP15CUV6+GHH2bJkiXW1x999BGPPPJINUYkNyolB9Xg+PHjWCwWWrVqVez2Vq1akZycTGJiYhVHVns5OTmxdOlSIiMjqVOnDrfccgsvvvgi+/fvr+7Qaq1HHnmE06dPs2nTJuu6jz76iKFDh+Lr61t9gdViI0eOZMuWLZw+fZrY2Fh+/vlnHnjggeoOS25ASg5qoCtXWjfSo6dvBH/96185f/48a9asoX///mzatInOnTuzdOnS6g6tVmrZsiU9evTgo48+AuDkyZP89NNPupKtRAEBAQwcOJDIyEiWLFnCwIEDCQgIqO6w5Aak5KAaNGvWDIPBwOHDh4vd/uuvv+Lr66t/1JXAzc2Nvn378vLLL7N161ZGjRpl0ycuFWv06NH8+9//Ji0tjSVLlhAWFkbv3r2rO6xa7ZFHHrFWyZSIyfVSclAN/P396du3L++//77NyGKA+Ph4li9fzr333qvKQRVo3bo1mZmZ1R1GrTVs2DAcHR1ZsWIFkZGRPPzww/q7rmQDBgzAZDJhMpno379/dYcjNyglB9VkwYIF5Obm0r9/f/773/8SFxdHVFQUffv2pX79+syYMaO6Q6xVLl++zB133MGyZcvYv38/MTExfPHFF8yZM4chQ4ZUd3i1lpeXF/feey8vvvgi58+f11z7KuDo6MiRI0c4cuQIjo6O1R2O3KCUHFST8PBwdu/eTdOmTbn33ntp2rQp48aN4/bbb2fbtm34+flVd4i1ipeXF926dWPevHncdttttG3blpdeeomxY8eyYMGC6g6vVhs9ejTJycn06dNH8+2riI+PDz4+PtUdhtzA9MhmERERsaHKgYiIiNhQciAiIiI2lByIiIiIDSUHIiIiYkPJgYiIiNhQciAiIiI2lByIiIiIDSUHIiIiYkPJgUgFmj59Oh07drS+HjVqFHfffXeVx3H69GkMBgPR0dHXbNOoUSPmz59f6mMuXbqUOnXqlDs2g8HA6tWry30cEak8Sg6k1hs1ahQGgwGDwYCzszNNmjTh2WefrZIHLr3zzjulfiR0aU7oIiJVwam6AxCpCgMGDGDJkiXk5eXx008/MWbMGDIzM1m4cGGRtnl5eTg7O1fI+xqNxgo5johIVVLlQOyCq6srwcHBhIaGMnz4cEaMGGEtbV/pCvjoo49o0qQJrq6uWCwWUlNTGTduHIGBgfj4+HDHHXewb98+m+O++eabBAUF4e3tzejRo8nJybHZ/vtuBbPZzOzZs2nWrBmurq40bNjQ+gTOxo0bA9CpUycMBgO9evWy7rdkyRJatWqFm5sbLVu25P3337d5n507d9KpUyfc3Nzo2rUrv/zyS5m/o7lz59KuXTs8PT0JDQ1lwoQJZGRkFGm3evVqmjdvjpubG3379iUuLs5m+zfffEOXLl1wc3OjSZMmvPrqq+Tn55c5HhGpPkoOxC65u7uTl5dnfX3ixAk+//xz/v3vf1vL+gMHDiQ+Pp7vvvuOPXv20LlzZ3r37k1SUhIAn3/+Oa+88gozZsxg9+7d1KtXr8hJ+/emTp3K7Nmzeemllzh8+DArVqwgKCgIKDzBA2zYsIELFy7w1VdfAbB48WKmTZvGjBkzOHLkCDNnzuSll14iMjISgMzMTAYNGkSLFi3Ys2cP06dP59lnny3zd+Lg4MA//vEPDh48SGRkJD/++CNTpkyxaZOVlcWMGTOIjIzk559/Ji0tjfvuu8+6/fvvv+eBBx5g0qRJHD58mEWLFrF06VI9glzkRmMRqeUeeughy5AhQ6yvd+zYYfH397cMGzbMYrFYLK+88orF2dnZkpCQYG3zww8/WHx8fCw5OTk2x2ratKll0aJFFovFYomIiLA8+uijNtu7detm6dChQ7HvnZaWZnF1dbUsXry42DhjYmIsgOWXX36xWR8aGmpZsWKFzbrXX3/dEhERYbFYLJZFixZZ/Pz8LJmZmdbtCxcuLPZYvxUWFmaZN2/eNbd//vnnFn9/f+vrJUuWWADL9u3breuOHDliASw7duywWCwWy6233mqZOXOmzXE++eQTS7169ayvAcuqVauu+b4iUv005kDswrfffouXlxf5+fnk5eUxZMgQ3n33Xev2sLAw6tata329Z88eMjIy8Pf3tzlOdnY2J0+eBODIkSM8+uijNtsjIiLYuHFjsTEcOXKE3NxcevfuXeq4ExMTiYuLY/To0YwdO9a6Pj8/3zqe4ciRI3To0AEPDw+bOMpq48aNzJw5k8OHD5OWlkZ+fj45OTlkZmbi6ekJgJOTE127drXu07JlS+rUqcORI0e4+eab2bNnD7t27bKpFBQUFJCTk0NWVpZNjCJScyk5ELtw++23s3DhQpydnQkJCSky4PDKye8Ks9lMvXr12LRpU5FjXe90Pnd39zLvYzabgcKuhW7dutlsc3R0BMBisVxXPL8VGxvLn//8Zx599FFef/11/Pz82LJlC6NHj7bpfoHCqYi/d2Wd2Wzm1VdfZejQoUXauLm5lTtOEakaSg7ELnh6etKsWbNSt+/cuTPx8fE4OTnRqFGjYtu0atWK7du38+CDD1rXbd++/ZrHDA8Px93dnR9++IExY8YU2e7i4gIUXmlfERQURP369Tl16hQjRowo9ritW7fmk08+ITs725qAlBRHcXbv3k1+fj5vv/02Dg6FQ5E+//zzIu3y8/PZvXs3N998MwBHjx4lJSWFli1bAoXf29GjR8v0XYtIzaPkQKQYffr0ISIigrvvvpvZs2fTokULzp8/z3fffcfdd99N165defLJJ3nooYfo2rUrf/rTn1i+fDmHDh2iSZMmxR7Tzc2N559/nilTpuDi4sItt9xCYmIihw4dYvTo0QQGBuLu7k5UVBQNGjTAzc0No9HI9OnTmTRpEj4+Ptx5553k5uaye/dukpOTefrppxk+fDjTpk1j9OjR/P3vf+f06dO89dZbZfq8TZs2JT8/n3fffZe77rqLn3/+mQ8++KBIO2dnZyZOnMg//vEPnJ2deeKJJ+jevbs1WXj55ZcZNGgQoaGh3HPPPTg4OLB//34OHDjAG2+8Ufb/ECJSLTRbQaQYBoOB7777jttuu41HHnmE5s2bc99993H69Gnr7IJ7772Xl19+meeff54uXboQGxvLY489VuJxX3rpJZ555hlefvllWrVqxb333ktCQgJQ2J//j3/8g0WLFhESEsKQIUMAGDNmDP/6179YunQp7dq1o2fPnixdutQ69dHLy4tvvvmGw4cP06lTJ6ZNm8bs2bPL9Hk7duzI3LlzmT17Nm3btmX58uXMmjWrSDsPDw+ef/55hg8fTkREBO7u7qxcudK6vX///nz77besX7+em266ie7duzN37lzCwsLKFI+IVC+DpSI6LEVERKTWUOVAREREbCg5EBERERtKDkRERMSGkgMRERGxoeRAREREbCg5EBERERtKDkRERMSGkgMRERGxoeRAREREbCg5EBERERtKDkRERMTG/wMWLC2xf1z0EAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm=confusion_matrix(ytest,ypred)\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=['O','S','V','M'])\n",
    "disp.plot()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "0e918adc-b6e6-4dea-8952-4f3e754fc6b3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7660999458676887, 0.0006646068560663582)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_f1=np.mean(f1)\n",
    "std_f1=np.std(f1)\n",
    "mean_f1, std_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "3a3873eb-21de-4d72-8b7e-c5a561cc0cf8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.3461486541458451, 0.00015229497530224365)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_precision=np.mean(precision)\n",
    "std_precision=np.std(precision)\n",
    "mean_precision, std_precision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "e082a635-8d99-45da-a7e1-2daaf10e89e8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7473228145155084, 0.0004767432377695068)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_recall=np.mean(recall)\n",
    "std_recall=np.std(recall)\n",
    "mean_recall, std_recall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "93881e3c-424c-46a7-bf82-bfcf9365f5a9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7473228145155084, 0.0004767432377695068)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_ba=np.mean(balanced_accuracy)\n",
    "std_ba=np.std(balanced_accuracy)\n",
    "mean_ba, std_ba"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "5e960be4-4921-4cb2-aad9-59ae538f158e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7306347917604934, 0.0007986981161830763)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_mc=np.mean(mc)\n",
    "std_mc=np.std(mc)\n",
    "mean_mc, std_mc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "323fef1f-c78d-4278-b410-090e1bb4b8fd",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02c9a5ff-eb13-4c62-bed4-0b80e3e07597",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.19"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}