grad_boost_with_resampling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "626df1d7",
   "metadata": {},
   "source": [
    "# bSMOTE: Borderline-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "68dfa32e",
   "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>AD_MCP</th>\n",
       "      <th>AD_PCT</th>\n",
       "      <th>AD_GCC</th>\n",
       "      <th>AD_BCC</th>\n",
       "      <th>AD_SCC</th>\n",
       "      <th>AD_FX</th>\n",
       "      <th>AD_CST_R</th>\n",
       "      <th>AD_CST_L</th>\n",
       "      <th>AD_ML_R</th>\n",
       "      <th>AD_ML_L</th>\n",
       "      <th>...</th>\n",
       "      <th>RD_FXST</th>\n",
       "      <th>RD_SLF</th>\n",
       "      <th>RD_SFO</th>\n",
       "      <th>RD_UNC</th>\n",
       "      <th>RD_TAP</th>\n",
       "      <th>RD_IC</th>\n",
       "      <th>RD_CC</th>\n",
       "      <th>RD_CR</th>\n",
       "      <th>RD_FULLWM</th>\n",
       "      <th>Research Group</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.001177</td>\n",
       "      <td>0.000974</td>\n",
       "      <td>0.001479</td>\n",
       "      <td>0.001474</td>\n",
       "      <td>0.001575</td>\n",
       "      <td>0.002257</td>\n",
       "      <td>0.001079</td>\n",
       "      <td>0.001081</td>\n",
       "      <td>0.001285</td>\n",
       "      <td>0.001287</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000662</td>\n",
       "      <td>0.000577</td>\n",
       "      <td>0.000551</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.001139</td>\n",
       "      <td>0.000491</td>\n",
       "      <td>0.000562</td>\n",
       "      <td>0.000638</td>\n",
       "      <td>0.000571</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.001213</td>\n",
       "      <td>0.001075</td>\n",
       "      <td>0.001567</td>\n",
       "      <td>0.001585</td>\n",
       "      <td>0.001686</td>\n",
       "      <td>0.002492</td>\n",
       "      <td>0.001046</td>\n",
       "      <td>0.001056</td>\n",
       "      <td>0.001149</td>\n",
       "      <td>0.001126</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000769</td>\n",
       "      <td>0.000694</td>\n",
       "      <td>0.000902</td>\n",
       "      <td>0.000717</td>\n",
       "      <td>0.001507</td>\n",
       "      <td>0.000626</td>\n",
       "      <td>0.000701</td>\n",
       "      <td>0.000851</td>\n",
       "      <td>0.000723</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.001101</td>\n",
       "      <td>0.000965</td>\n",
       "      <td>0.001440</td>\n",
       "      <td>0.001433</td>\n",
       "      <td>0.001456</td>\n",
       "      <td>0.002379</td>\n",
       "      <td>0.000985</td>\n",
       "      <td>0.000990</td>\n",
       "      <td>0.001118</td>\n",
       "      <td>0.001113</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000614</td>\n",
       "      <td>0.000546</td>\n",
       "      <td>0.000562</td>\n",
       "      <td>0.000553</td>\n",
       "      <td>0.000950</td>\n",
       "      <td>0.000448</td>\n",
       "      <td>0.000512</td>\n",
       "      <td>0.000587</td>\n",
       "      <td>0.000531</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.001053</td>\n",
       "      <td>0.000904</td>\n",
       "      <td>0.001433</td>\n",
       "      <td>0.001470</td>\n",
       "      <td>0.001450</td>\n",
       "      <td>0.002421</td>\n",
       "      <td>0.000966</td>\n",
       "      <td>0.000977</td>\n",
       "      <td>0.001047</td>\n",
       "      <td>0.001041</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000607</td>\n",
       "      <td>0.000535</td>\n",
       "      <td>0.000551</td>\n",
       "      <td>0.000530</td>\n",
       "      <td>0.000862</td>\n",
       "      <td>0.000437</td>\n",
       "      <td>0.000509</td>\n",
       "      <td>0.000573</td>\n",
       "      <td>0.000522</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.001131</td>\n",
       "      <td>0.000974</td>\n",
       "      <td>0.001469</td>\n",
       "      <td>0.001498</td>\n",
       "      <td>0.001531</td>\n",
       "      <td>0.002291</td>\n",
       "      <td>0.000994</td>\n",
       "      <td>0.001017</td>\n",
       "      <td>0.001128</td>\n",
       "      <td>0.001105</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000621</td>\n",
       "      <td>0.000559</td>\n",
       "      <td>0.000586</td>\n",
       "      <td>0.000558</td>\n",
       "      <td>0.000901</td>\n",
       "      <td>0.000458</td>\n",
       "      <td>0.000527</td>\n",
       "      <td>0.000605</td>\n",
       "      <td>0.000545</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>963</th>\n",
       "      <td>0.001117</td>\n",
       "      <td>0.001011</td>\n",
       "      <td>0.001574</td>\n",
       "      <td>0.001646</td>\n",
       "      <td>0.001712</td>\n",
       "      <td>0.002214</td>\n",
       "      <td>0.001094</td>\n",
       "      <td>0.001118</td>\n",
       "      <td>0.001228</td>\n",
       "      <td>0.001248</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000734</td>\n",
       "      <td>0.000647</td>\n",
       "      <td>0.000913</td>\n",
       "      <td>0.000599</td>\n",
       "      <td>0.001623</td>\n",
       "      <td>0.000538</td>\n",
       "      <td>0.000680</td>\n",
       "      <td>0.000744</td>\n",
       "      <td>0.000652</td>\n",
       "      <td>AD</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>964</th>\n",
       "      <td>0.001181</td>\n",
       "      <td>0.001002</td>\n",
       "      <td>0.001423</td>\n",
       "      <td>0.001505</td>\n",
       "      <td>0.001530</td>\n",
       "      <td>0.002266</td>\n",
       "      <td>0.001088</td>\n",
       "      <td>0.001103</td>\n",
       "      <td>0.001224</td>\n",
       "      <td>0.001203</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000584</td>\n",
       "      <td>0.000593</td>\n",
       "      <td>0.000624</td>\n",
       "      <td>0.000619</td>\n",
       "      <td>0.000878</td>\n",
       "      <td>0.000487</td>\n",
       "      <td>0.000567</td>\n",
       "      <td>0.000618</td>\n",
       "      <td>0.000571</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>965</th>\n",
       "      <td>0.001126</td>\n",
       "      <td>0.000949</td>\n",
       "      <td>0.001542</td>\n",
       "      <td>0.001747</td>\n",
       "      <td>0.001642</td>\n",
       "      <td>0.002565</td>\n",
       "      <td>0.000978</td>\n",
       "      <td>0.001052</td>\n",
       "      <td>0.001162</td>\n",
       "      <td>0.001190</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000577</td>\n",
       "      <td>0.000589</td>\n",
       "      <td>0.000835</td>\n",
       "      <td>0.000534</td>\n",
       "      <td>0.001478</td>\n",
       "      <td>0.000440</td>\n",
       "      <td>0.000694</td>\n",
       "      <td>0.000659</td>\n",
       "      <td>0.000597</td>\n",
       "      <td>AD</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>966</th>\n",
       "      <td>0.001090</td>\n",
       "      <td>0.000976</td>\n",
       "      <td>0.001354</td>\n",
       "      <td>0.001421</td>\n",
       "      <td>0.001427</td>\n",
       "      <td>0.001928</td>\n",
       "      <td>0.001018</td>\n",
       "      <td>0.001007</td>\n",
       "      <td>0.001241</td>\n",
       "      <td>0.001188</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000560</td>\n",
       "      <td>0.000517</td>\n",
       "      <td>0.000514</td>\n",
       "      <td>0.000566</td>\n",
       "      <td>0.000729</td>\n",
       "      <td>0.000441</td>\n",
       "      <td>0.000469</td>\n",
       "      <td>0.000534</td>\n",
       "      <td>0.000509</td>\n",
       "      <td>CN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>967</th>\n",
       "      <td>0.001097</td>\n",
       "      <td>0.000984</td>\n",
       "      <td>0.001477</td>\n",
       "      <td>0.001419</td>\n",
       "      <td>0.001502</td>\n",
       "      <td>0.002556</td>\n",
       "      <td>0.001063</td>\n",
       "      <td>0.001147</td>\n",
       "      <td>0.001268</td>\n",
       "      <td>0.001199</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000601</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.000766</td>\n",
       "      <td>0.000605</td>\n",
       "      <td>0.001115</td>\n",
       "      <td>0.000446</td>\n",
       "      <td>0.000598</td>\n",
       "      <td>0.000608</td>\n",
       "      <td>0.000556</td>\n",
       "      <td>MCI</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>933 rows × 293 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       AD_MCP    AD_PCT    AD_GCC    AD_BCC    AD_SCC     AD_FX  AD_CST_R  \\\n",
       "0    0.001177  0.000974  0.001479  0.001474  0.001575  0.002257  0.001079   \n",
       "2    0.001213  0.001075  0.001567  0.001585  0.001686  0.002492  0.001046   \n",
       "3    0.001101  0.000965  0.001440  0.001433  0.001456  0.002379  0.000985   \n",
       "4    0.001053  0.000904  0.001433  0.001470  0.001450  0.002421  0.000966   \n",
       "5    0.001131  0.000974  0.001469  0.001498  0.001531  0.002291  0.000994   \n",
       "..        ...       ...       ...       ...       ...       ...       ...   \n",
       "963  0.001117  0.001011  0.001574  0.001646  0.001712  0.002214  0.001094   \n",
       "964  0.001181  0.001002  0.001423  0.001505  0.001530  0.002266  0.001088   \n",
       "965  0.001126  0.000949  0.001542  0.001747  0.001642  0.002565  0.000978   \n",
       "966  0.001090  0.000976  0.001354  0.001421  0.001427  0.001928  0.001018   \n",
       "967  0.001097  0.000984  0.001477  0.001419  0.001502  0.002556  0.001063   \n",
       "\n",
       "     AD_CST_L   AD_ML_R   AD_ML_L  ...   RD_FXST    RD_SLF    RD_SFO  \\\n",
       "0    0.001081  0.001285  0.001287  ...  0.000662  0.000577  0.000551   \n",
       "2    0.001056  0.001149  0.001126  ...  0.000769  0.000694  0.000902   \n",
       "3    0.000990  0.001118  0.001113  ...  0.000614  0.000546  0.000562   \n",
       "4    0.000977  0.001047  0.001041  ...  0.000607  0.000535  0.000551   \n",
       "5    0.001017  0.001128  0.001105  ...  0.000621  0.000559  0.000586   \n",
       "..        ...       ...       ...  ...       ...       ...       ...   \n",
       "963  0.001118  0.001228  0.001248  ...  0.000734  0.000647  0.000913   \n",
       "964  0.001103  0.001224  0.001203  ...  0.000584  0.000593  0.000624   \n",
       "965  0.001052  0.001162  0.001190  ...  0.000577  0.000589  0.000835   \n",
       "966  0.001007  0.001241  0.001188  ...  0.000560  0.000517  0.000514   \n",
       "967  0.001147  0.001268  0.001199  ...  0.000601  0.000537  0.000766   \n",
       "\n",
       "       RD_UNC    RD_TAP     RD_IC     RD_CC     RD_CR  RD_FULLWM  \\\n",
       "0    0.000600  0.001139  0.000491  0.000562  0.000638   0.000571   \n",
       "2    0.000717  0.001507  0.000626  0.000701  0.000851   0.000723   \n",
       "3    0.000553  0.000950  0.000448  0.000512  0.000587   0.000531   \n",
       "4    0.000530  0.000862  0.000437  0.000509  0.000573   0.000522   \n",
       "5    0.000558  0.000901  0.000458  0.000527  0.000605   0.000545   \n",
       "..        ...       ...       ...       ...       ...        ...   \n",
       "963  0.000599  0.001623  0.000538  0.000680  0.000744   0.000652   \n",
       "964  0.000619  0.000878  0.000487  0.000567  0.000618   0.000571   \n",
       "965  0.000534  0.001478  0.000440  0.000694  0.000659   0.000597   \n",
       "966  0.000566  0.000729  0.000441  0.000469  0.000534   0.000509   \n",
       "967  0.000605  0.001115  0.000446  0.000598  0.000608   0.000556   \n",
       "\n",
       "     Research Group  \n",
       "0                CN  \n",
       "2                CN  \n",
       "3                CN  \n",
       "4                CN  \n",
       "5                CN  \n",
       "..              ...  \n",
       "963              AD  \n",
       "964              CN  \n",
       "965              AD  \n",
       "966              CN  \n",
       "967             MCI  \n",
       "\n",
       "[933 rows x 293 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "df = pd.read_csv('DTI_with_group.csv')\n",
    "# remove unnecessary columns + confounding\n",
    "df = df.drop(['IMAGEUID', 'COLPROT', 'RID', 'VISCODE', \n",
    "                          'VISCODE2', 'EXAMDATE', 'VERSION', 'MANUFACTURER',\n",
    "                         'RUNDATE', 'STATUS', 'QC', 'update_stamp', 'Age', 'VOLUMES', 'Sex'], axis=1)\n",
    "# remove rows with empty values\n",
    "df = df.dropna()\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "403d3a61",
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import BorderlineSMOTE as bSMOTE\n",
    "X = df.iloc[:, df.columns != 'Research Group'].values\n",
    "y = df.iloc[:, -1].values\n",
    "X_resampled, y_resampled = bSMOTE(kind='borderline-1').fit_resample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "774f7ac5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from numpy import mean\n",
    "from numpy import std\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report, confusion_matrix, accuracy_score\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size = 0.33, random_state = 8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "03677028",
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f082b1bf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate:  0.05\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.905\n",
      "Learning rate:  0.075\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.906\n",
      "Learning rate:  0.1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.906\n",
      "Learning rate:  0.25\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.911\n",
      "Learning rate:  0.5\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.904\n",
      "Learning rate:  0.75\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.889\n",
      "Learning rate:  1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.884\n"
     ]
    }
   ],
   "source": [
    "lr_list = [0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1]\n",
    "\n",
    "for learning_rate in lr_list:\n",
    "    gb_clf = GradientBoostingClassifier(n_estimators=2000, learning_rate=learning_rate, max_depth=5, random_state=8)\n",
    "    gb_clf.fit(X_train, y_train)\n",
    "\n",
    "    print(\"Learning rate: \", learning_rate)\n",
    "    print(\"Accuracy score (training): {0:.3f}\".format(gb_clf.score(X_train, y_train)))\n",
    "    print(\"Accuracy score (validation): {0:.3f}\".format(gb_clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "58ec9121",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[147   0   0   0   3   0]\n",
      " [  4 113  10   2  30   1]\n",
      " [  2   8 139   3   4   1]\n",
      " [  1   1   0 164   1   0]\n",
      " [  1  12   2   1 152   3]\n",
      " [  1   1   1   0   0 167]]\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          AD       0.94      0.98      0.96       150\n",
      "          CN       0.84      0.71      0.77       160\n",
      "        EMCI       0.91      0.89      0.90       157\n",
      "        LMCI       0.96      0.98      0.97       167\n",
      "         MCI       0.80      0.89      0.84       171\n",
      "         SMC       0.97      0.98      0.98       170\n",
      "\n",
      "    accuracy                           0.90       975\n",
      "   macro avg       0.90      0.90      0.90       975\n",
      "weighted avg       0.90      0.90      0.90       975\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gb_clf2 = GradientBoostingClassifier(n_estimators=2000, learning_rate=0.05, max_depth=5, random_state=8)\n",
    "gb_clf2.fit(X_train, y_train)\n",
    "predictions = gb_clf2.predict(X_test)\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, predictions))\n",
    "\n",
    "print(\"Classification Report\")\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b4198f06",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, predictions)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(8, 8))\n",
    "ax.imshow(cm)\n",
    "ax.grid(False)\n",
    "ax.set_xlabel('Predicted outputs', fontsize=12, color='black')\n",
    "ax.set_ylabel('Actual outputs', fontsize=12, color='black')\n",
    "ax.xaxis.set(ticks=range(6))\n",
    "ax.yaxis.set(ticks=range(6))\n",
    "for i in range(6):\n",
    "    for j in range(6):\n",
    "        ax.text(j, i, cm[i, j], ha='center', va='center', color='red')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "4eb4f68b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "# Normalise\n",
    "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show(block=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "582164bd",
   "metadata": {},
   "source": [
    "### Combined MCI groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "71d7768f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('DTI_with_group_comb.csv')\n",
    "# remove unnecessary columns + confounding\n",
    "df = df.drop(['IMAGEUID', 'COLPROT', 'RID', 'VISCODE', \n",
    "                          'VISCODE2', 'EXAMDATE', 'VERSION', 'MANUFACTURER',\n",
    "                         'RUNDATE', 'STATUS', 'QC', 'update_stamp', 'Age', 'VOLUMES', 'Sex'], axis=1)\n",
    "# remove rows with empty values\n",
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "aa99fca0",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.iloc[:, df.columns != 'Research Group'].values\n",
    "y = df.iloc[:, -1].values\n",
    "X_resampled, y_resampled = bSMOTE(kind='borderline-1').fit_resample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "352c0193",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size = 0.33, random_state = 11)\n",
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ec224ba6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate:  0.05\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.794\n",
      "Learning rate:  0.075\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.794\n",
      "Learning rate:  0.1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.794\n",
      "Learning rate:  0.25\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.812\n",
      "Learning rate:  0.5\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.809\n",
      "Learning rate:  0.75\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.812\n",
      "Learning rate:  1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.792\n"
     ]
    }
   ],
   "source": [
    "lr_list = [0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1]\n",
    "\n",
    "for learning_rate in lr_list:\n",
    "    gb_clf = GradientBoostingClassifier(n_estimators=2000, learning_rate=learning_rate, max_depth=5, random_state=11)\n",
    "    gb_clf.fit(X_train, y_train)\n",
    "\n",
    "    print(\"Learning rate: \", learning_rate)\n",
    "    print(\"Accuracy score (training): {0:.3f}\".format(gb_clf.score(X_train, y_train)))\n",
    "    print(\"Accuracy score (validation): {0:.3f}\".format(gb_clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "fa15f0e2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[180   3   5]\n",
      " [  4 135  43]\n",
      " [  8  42 129]]\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          AD       0.94      0.96      0.95       188\n",
      "          CN       0.75      0.74      0.75       182\n",
      "         MCI       0.73      0.72      0.72       179\n",
      "\n",
      "    accuracy                           0.81       549\n",
      "   macro avg       0.81      0.81      0.81       549\n",
      "weighted avg       0.81      0.81      0.81       549\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gb_clf2 = GradientBoostingClassifier(n_estimators=2000, learning_rate=0.5, max_depth=5, random_state=11)\n",
    "gb_clf2.fit(X_train, y_train)\n",
    "predictions = gb_clf2.predict(X_test)\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, predictions))\n",
    "\n",
    "print(\"Classification Report\")\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7233f2fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, predictions)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(8, 8))\n",
    "ax.imshow(cm)\n",
    "ax.grid(False)\n",
    "ax.set_xlabel('Predicted outputs', fontsize=12, color='black')\n",
    "ax.set_ylabel('Actual outputs', fontsize=12, color='black')\n",
    "ax.xaxis.set(ticks=range(3))\n",
    "ax.yaxis.set(ticks=range(3))\n",
    "for i in range(3):\n",
    "    for j in range(3):\n",
    "        ax.text(j, i, cm[i, j], ha='center', va='center', color='white')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "4e748280",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "# Normalise\n",
    "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show(block=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "020fd47a",
   "metadata": {},
   "source": [
    "# bSMOTE with SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "13fbcb85",
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SVMSMOTE\n",
    "df = pd.read_csv('DTI_with_group.csv')\n",
    "# remove unnecessary columns + confounding\n",
    "df = df.drop(['IMAGEUID', 'COLPROT', 'RID', 'VISCODE', \n",
    "                          'VISCODE2', 'EXAMDATE', 'VERSION', 'MANUFACTURER',\n",
    "                         'RUNDATE', 'STATUS', 'QC', 'update_stamp', 'Age', 'VOLUMES', 'Sex'], axis=1)\n",
    "# remove rows with empty values\n",
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "898445dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.iloc[:, df.columns != 'Research Group'].values\n",
    "y = df.iloc[:, -1].values\n",
    "X_resampled, y_resampled = SVMSMOTE().fit_resample(X, y)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3, random_state=23)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ad16792b",
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "90bfe966",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate:  0.05\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.838\n",
      "Learning rate:  0.075\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.836\n",
      "Learning rate:  0.1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.834\n",
      "Learning rate:  0.25\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.828\n",
      "Learning rate:  0.5\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.840\n",
      "Learning rate:  0.75\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.826\n",
      "Learning rate:  1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.806\n"
     ]
    }
   ],
   "source": [
    "lr_list = [0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1]\n",
    "\n",
    "for learning_rate in lr_list:\n",
    "    gb_clf = GradientBoostingClassifier(n_estimators=2000, learning_rate=learning_rate, max_depth=5, random_state=23)\n",
    "    gb_clf.fit(X_train, y_train)\n",
    "\n",
    "    print(\"Learning rate: \", learning_rate)\n",
    "    print(\"Accuracy score (training): {0:.3f}\".format(gb_clf.score(X_train, y_train)))\n",
    "    print(\"Accuracy score (validation): {0:.3f}\".format(gb_clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "46207a52",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[ 65   4   0   1   2   0]\n",
      " [  2 113  10   0  21   6]\n",
      " [  0   8  69   3   4   0]\n",
      " [  0   1   0  63   5   0]\n",
      " [  1  19   2   0 126   1]\n",
      " [  0   3   2   0   1  60]]\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          AD       0.96      0.90      0.93        72\n",
      "          CN       0.76      0.74      0.75       152\n",
      "        EMCI       0.83      0.82      0.83        84\n",
      "        LMCI       0.94      0.91      0.93        69\n",
      "         MCI       0.79      0.85      0.82       149\n",
      "         SMC       0.90      0.91      0.90        66\n",
      "\n",
      "    accuracy                           0.84       592\n",
      "   macro avg       0.86      0.86      0.86       592\n",
      "weighted avg       0.84      0.84      0.84       592\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gb_clf2 = GradientBoostingClassifier(n_estimators=2000, learning_rate=0.05, max_depth=5, random_state=23)\n",
    "gb_clf2.fit(X_train, y_train)\n",
    "predictions = gb_clf2.predict(X_test)\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, predictions))\n",
    "\n",
    "print(\"Classification Report\")\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "1e8efc5a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, predictions)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(8, 8))\n",
    "ax.imshow(cm)\n",
    "ax.grid(False)\n",
    "ax.set_xlabel('Predicted outputs', fontsize=12, color='black')\n",
    "ax.set_ylabel('Actual outputs', fontsize=12, color='black')\n",
    "ax.xaxis.set(ticks=range(6))\n",
    "ax.yaxis.set(ticks=range(6))\n",
    "for i in range(6):\n",
    "    for j in range(6):\n",
    "        ax.text(j, i, cm[i, j], ha='center', va='center', color='white')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d0b37cd9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxgAAANBCAYAAABj/BhbAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAACFyklEQVR4nOzdd3gU5RbH8d+GhEAiLaRQRNFLE5AWIEFRBBFQpAgiVhAQ6VVQehcUkSJdmtKbBZAmhN4hgYTem0IaVUFDwu79A112yQYIO7Cb+P3cZ5/n8uadyRnGHfbsOfOOyWKxWAQAAAAABvBwdQAAAAAA0g8SDAAAAACGIcEAAAAAYBgSDAAAAACGIcEAAAAAYBgSDAAAAACGIcEAAAAAYBgSDAAAAACGIcEAAAAAYBhPVwfwMCTGHHZ1CLiLfIXrujoEpCD++lVXh4AUZPDg+yB35uWRLv85TRf+Trrh6hCQgqQbv7s6hBQlxp9wdQgp8vJ/2tUh3BP/YgEAAAAwDAkGAAAAAMNQ0wUAAABsmW+6OoI0jQoGAAAAAMOQYAAAAAAwDC1SAAAAgC2L2dURpGlUMAAAAAAYhgQDAAAAgGFokQIAAABsmWmRcgYVDAAAAACGIcEAAAAAYBhapAAAAAAbFlaRcgoVDAAAAACGIcEAAAAAYBhapAAAAABbrCLlFCoYAAAAAAxDggEAAADAMLRIAQAAALZYRcopVDAAAAAAGIYEAwAAAIBhaJECAAAAbJlvujqCNI0KBgAAAADDkGAAAAAAMAwtUgAAAIAtVpFyChUMAAAAAIYhwQAAAABgGFqkAAAAAFtmWqScQQUDAAAAgGFIMAAAAAAYhhYpAAAAwIaFVaScQgUDAAAAgGFIMAAAAAAYhhYpAAAAwBarSDmFCgYAAAAAw5BgAAAAADAMLVIAAACALVaRcgoVDAAAAACGIcEAAAAAYBhapAAAAABb5puujiBNo4IBAAAAwDAkGAAAAAAMQ4sUAAAAYItVpJxCBQMAAACAYUgwAAAAABiGFikAAADAlpkWKWdQwQAAAABgGBIMAAAAAIahRQoAAACwxSpSTqGCAQAAAMAwJBgAAAAADEOLFAAAAGCLVaScQgUDAAAAgGFIMAAAAAAYhhYpAAAAwIbFctPVIaRpVDAAAAAAGIYEAwAAAIBhaJECAAAAbPGgPadQwXjE5vy4VNXe+khlqtbXOy26aO+BIynOTUxK0vjv5qrG2x+rTNX6qtekvTZtD3dqn7i7Jh+9q51Rq3Uqeo+WrZ6r0mWevev8WnWqa+OOpToVvUdrNy/Sy6+8mGxOwUJP6/s5Y3Xk9A6d+D1cK9bMV97Hcz+sQ0i3WrVsrGNHtunPq8e1ZdMSlStb6q7z69d/Xfv2rtefV49rd8RqvVqjSrI5/fp20dnTEfrjyjGtXD5XBQo89ZCiT99atmisw4e36Mrlo9q4YbHK3uPc1KtXU1GRa3Xl8lGF71qlGtUr2/28Tp0aWvrLLJ37PUoJf59ViRJFH2L06d/HLT7Q/oMbFX/xkNau/0nBZUvedf4bb7ymiN2rFX/xkLbvWK5q1V+y+3mPnh0UsXu1YuL26+zve7TklxkqW67UwzuAdIzrGtIrEoxHaHnYRg0dO0WtPnxbCyaPUOEC+dWiS19duHTZ4fzRk2ZqweIV6tHhYy2aPlZv1amhDj2H6OCR4w+8T6Sszhuvqt/nn+nrL8eqWqX62r/vsOb8OEn+/n4O55ctX0rjpwzTnBk/6JUX62n5sjBNmzVaRZ4paJ3zZP58WrRilo4dOal6tRqr8vN1Nfyr8Ur4O+FRHVa60KBBbQ37qq8GDhquciE1FBl1QMuWzlJAQE6H8yuEltWsGWM1bdoclS1fXYsXr9QPC6eoWLHC1jldu7RW2zZN1bptNz1XsZauXb+uZb/Mkre396M6rHThzTdraejQ3vr885EKCX1Ne/ce0C9LZqR4bkJDgzVj+hh9991chYS8qsVLVmrBgskqWvT2ufH19dHmLTvUs9fgR3UY6Vb9+jU15IueGjJ4lCo+97r27T2onxd9n+L5CQkpo2nfj9L338/X8xVq6pdfVmnuvIkqWrSQdc7RoyfVuXNfhZSroWpVG+j0md+1aPH3KV4r4RjXNaRnJovFYnF1EEZLjDns6hAceqdFFxUvUkA9O7WUJJnNZlV9s6nerfe6Pnr/zWTzK7/xoT7+oIHeqVfTOtax1xB5e2fUl70/eaB9uoN8heu6OgSHlq2eqz0R+9Tj00GSJJPJpIj9azXl25kaM3JysvkTpw6Xj09mffB2K+vY0lVztW/vQX3Wub8kacKUr5WYlKR2LT57NAfhpPjrV10dgkNbNi3Rzl2R6tCxl6Rb5+bUiZ0aO26ahn41Ntn82bPGy9fHR3XeaGwd27xxifZE7lebtt0kSWdPR2jEyIkaPmKiJClr1iw699seNf2ok+bPX/wIjip1Mni45/dBGzcsVnh4pDp26i3p1rk5fmyHxo2fpmHDxiWbP3PGOPn6ZtYb9ZpYxzasX6SoqP1q266H3dwnn3xcRw5vVbny1RUVdeDhHoiTvDzcs+N47fqfFBEepU8695V06/wcPrpFE8Z/r+FfT0g2//vpo+Xjm1kN6n9kHVuz7kftjTqgDu17OfwdWbI8pvMxe/X6a+9p3botD+dAnPB30g1Xh+AQ1zUp6cbvrg4hRX9HuN/f178ylant6hDuyT3/xUqHEhMTdeDIMYXalD89PDwUGlxSkfsPOdzmRmKiMmb0shvz9s6o3XsPPvA+4ZiXl5dKlCqmDeu3WscsFos2rt+qsuVLOdwmuFxJu/mStG7NJut8k8mkqtUq6cSxU5rzwyTtO7pJy1bPVY2aLz+sw0iXvLy8VKZMCYWt2Wgds1gsCluzSaGhwQ63CQ0JtpsvSb+uWmed/9RTTyh37iCFrdlk/fnVq39ox47dCg1xvE8kd+vcPKs1Nn+PFotFa9ZuTPHvMSS0jN18SVq1er1C+Hs3nJeXl0qXLq61a+3Pz9o1m1U+pIzDbcqHlNbaNZvtxsJWb1D58o7ne3l5qUnTd3T58lXt/effJtwb1zWkdy5NMOLj4zV06FC98cYbqlChgipUqKA33nhDX331leLi4lwZmuEuXbmqmzfNypkju914Tr/sir942eE2z5cvrenzF+n02XMym83asnO3wjZsVdyFiw+8TzjmlzO7PD09FRd7wW48LvaCAgP9HW4TGOSvuNj4FOf7B+TUY1l81a7jR1obtkkN632kZb+s1tQZ36jC8+UezoGkQ/7+fvL09FRsjP3fdWxsnHIFBTjcJleuAMXE2l9DYmLirfNzBQX+M3bHnNh45coVaFTo6d6/5+bOv+vYmHgFpXRuggIUc8f75m7z8eBy+udI4b2T8t93UFBAsuuao/k1Xq2i6Nh9unDpkNq2a6ratT7QhQuXjD2AdIzrGtI7l9V0d+7cqerVq8vHx0dVq1ZVoUK3+jtjYmL0zTff6IsvvtDKlStVtmzZu+4nISFBCQn2/eweCTfk7Z3xocX+qHRr31z9ho5RrQ9ay2SS8uXJrbqvVtVPy1a7OjTcBw8PkyRpxbI1+nbc95Kk/XsPqVxIaTVq0lBbN+90ZXgA8MA2rN+q50JrKmfOHPqw6duaPmOMKld6Q3FxF+69MZAWsIqUU1yWYLRr104NGjTQhAkTZDKZ7H5msVjUsmVLtWvXTlu3bk1hD7cMGTJE/fv3txvr9Ukb9enazvCYnZEjW1ZlyOCR7ObrCxcvy98vu8Nt/LJn0zeDeyoh4YYuX/1Dgf5+GjHhez2eJ+iB9wnHLl64rKSkJAUE2t9cFxCYU7F3fJv3r9iYeAXcUd2wnX/xwmUlJibqyOHjdnOOHj6h8qGO2w2QXHz8RSUlJSkwyP7vOjAwQNExjiud0dFxCgq0/xYwKMjfOj86JvafsQBFR8fenhPorz2R+40MP13799zc+XcdGOSf7FvUf0XHxCnojvfN3ebjwV2Iv5TCeyflv++YmLhk1zVH869f/0snTpzWiROntXPnHu2JWqNGjd/S18PGG3sQ6RTXNaR3LmuRioyMVKdOnZIlF9Kt3vVOnTppz54999xP9+7ddeXKFbvXZ+1bPISInePl5aWihQpoe3ikdcxsNmt7RJRKFity1229vTMqKCCnkm7e1KoNW1S5YojT+4S9xMRERe3ZrxcqhVrHTCaTKr4Yql079jjcJnxnpN18SXrxpees8xMTE7UnYp/+V9B+icCnC+TXb2fPGRp/epaYmKiIiChVqVzROmYymVSlckVt25Z82WZJ2rY9XFWqVLQbq/ryi9b5J0+e0fnzMXb7zJLlMZUvX1rbHCwFDcdunZu9qlz5eeuYyWRS5Zcqpvj3uH1bhN18SXq5ygvazt+74RITE7V79z699JL9+Xmp8nPasT3C4TY7tu/WS3ecn8pVKmrHDsfz/+Xh4ZEuOgceFa5rSO9cVsHIlSuXduzYoSJFHH8Q3rFjh4KCgu65H29v72TLryX+5Z4XuUZv1VHPISNVrHABFX+mkGYuWKy//vpbdV+7ddNv989HKNDfT51a3FohIurAYcXEXVCRgk8rNu6Cxk2bI4vZoqbv1LvvfeL+TRz7vUaNH6LI3fu0O3yvmrdqJB/fzJo76ydJ0ugJX+j8uRgNHjBCkjRpwnT9tHS6Wrb9UKtXrlfd+q+pZOli6tqxr3Wf40ZP1cSpX2vb5l3avHG7qlStqGo1XlK91xs7jAGOjRg1SdOmjFB4RJR27tyt9u2ay9c3s777fp4kadrUUTp37rx69vpCkjR69BStCVuoTh1baNny1Wr4Vh0FB5dQy9afWvf5zejJ6tG9vY4eO6FTp86qf7+uOncuRosWrXTJMaZVo76ZpCmThys8Ikq7du5Ru3bN5OubWdOnz5ckTZkyQufORat37y8lSWPGTtHqVQvUscPHWr48TA3eqq3g4BJq3aabdZ85cmRXvnx5lCf3rX8DChX6n6Rb365T6UidMd9M1sRJXysiIkrhuyLVpm1T+fj4aOaMhZKkbyd9rXPnotWv71eSpHFjp2nFr3PVrv1HWrlijd5sUEtlyjyr9m1vrfDl45NZXT9ro2W/rFZ0dJxy+ufQxy0+UJ48ufTTj8tcdpxpEdc1N2e+6eoI0jSXJRhdunTRxx9/rPDwcL388svWZCImJkZhYWGaNGmShg0b5qrwHopXX35Bly5f0ZipsxV/8ZKKFHhaE4b1k79fDknS+Zg4edhUdBJuJGr05Fn67Xy0fDJn0guhZTWkVydlzfLYfe8T92/RT8uV0z+HPu3RXgGB/tq/96Deqf+x4v/pKc77eG6Zzbd7Mnft2KPWH3XVZ706qHvvTjp5/LSavNdOhw4etc5Z/stqfda5v9p1+liDvuyh48dOqlmjDtqx7e7fBsLeggWLFeDvp359uihXrgBFRu5Xzdfft7ajPZEvj9252bptl95v1FYD+n+qQQM/09FjJ1X/zWbav//2EtZfDRsnX18fTRg3VNmzZ9XmzTtVs9b7ye7pwt0tXLhEAf5+6tPnE+UKClBk5AHVqv2B9dzky5dXZvPt1dC3bQtXo8bt1L9fVw0Y8KmOHTulBg0+0oEDt8/N66+/osmThlv/PGvmreVuBw4arkGDRjyiI0sffvhhqfwDcqpX784KCvJXVNRBvVH3Q5vzY//e2b49Qk0/7KjefT9Rv/5ddPzYKb3dsIUO/PMA15s3b6pwof/pvTn1lTNnDl28eFnh4VGq9spbOmhz7cO9cV1DeubS52DMmzdPI0aMUHh4uG7evJUpZsiQQcHBwercubPeeuutB9qvuz4HA7e463Mw4L7PwYD7PgcDt7jrczDgvs/BgJs/B2PnD64OIUWZytV3dQj35NIrYsOGDdWwYUMlJiYqPv5Wxu7v7y8vL697bAkAAAA8JKwi5RS3+MrFy8tLuXPndnUYAAAAAJxEzR0AAACAYdyiggEAAAC4DTMtUs6gggEAAADAMCQYAAAAAAxDixQAAABgi1WknEIFAwAAAIBhSDAAAAAAGIYWKQAAAMAWq0g5hQoGAAAAAMOQYAAAAAAwDC1SAAAAgC1apJxCBQMAAACAYUgwAAAAABiGFikAAADAhsVy09UhpGlUMAAAAAAYhgQDAAAAgGFokQIAAABssYqUU6hgAAAAADAMCQYAAAAAw9AiBQAAANiy0CLlDCoYAAAAQDo1duxY5c+fX5kyZVJISIh27Nhx1/kjR45U4cKFlTlzZuXLl0+dOnXS33//narfSYIBAAAApEPz5s1T586d1bdvX0VERKhkyZKqXr26YmNjHc6fPXu2unXrpr59++rgwYOaMmWK5s2bpx49eqTq95JgAAAAALbMZvd9pcLw4cPVvHlzNWnSREWLFtWECRPk4+OjqVOnOpy/ZcsWPf/883r33XeVP39+VatWTe+88849qx53IsEAAAAA0oiEhARdvXrV7pWQkJBs3o0bNxQeHq6qVataxzw8PFS1alVt3brV4b6fe+45hYeHWxOKEydOaNmyZXrttddSFSMJBgAAAJBGDBkyRNmyZbN7DRkyJNm8+Ph43bx5U0FBQXbjQUFBio6Odrjvd999VwMGDFDFihXl5eWl//3vf3rppZdokQIAAACcYjG77at79+66cuWK3at79+6GHPa6des0ePBgjRs3ThEREfrxxx+1dOlSDRw4MFX7YZlaAAAAII3w9vaWt7f3Pef5+/srQ4YMiomJsRuPiYlRrly5HG7Tu3dvffDBB/roo48kSc8++6yuXbumjz/+WD179pSHx/3VJqhgAAAAAOlMxowZFRwcrLCwMOuY2WxWWFiYKlSo4HCb69evJ0siMmTIIEmyWCz3/bupYAAAAAC2Urlak7vq3LmzGjdurLJly6p8+fIaOXKkrl27piZNmkiSGjVqpLx581rv4ahVq5aGDx+u0qVLKyQkRMeOHVPv3r1Vq1Yta6JxP0gwAAAAgHSoYcOGiouLU58+fRQdHa1SpUppxYoV1hu/z5w5Y1ex6NWrl0wmk3r16qXff/9dAQEBqlWrlj7//PNU/V6TJTX1jjQiMeawq0PAXeQrXNfVISAF8devujoEpCDDffa9wjW8PPi+zl39nXTD1SEgBUk3fnd1CCn669dxrg4hRZmrtXZ1CPfEFREAAACwZUkfLVKuwldiAAAAAAxDggEAAADAMLRIAQAAALbSySpSrkIFAwAAAIBhSDAAAAAAGIYWKQAAAMAWLVJOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYIsH7TmFCgYAAAAAw5BgAAAAADAMLVIAAACALVaRcgoVDAAAAACGIcEAAAAAYBhapAAAAABbrCLlFCoYAAAAAAxDggEAAADAMLRIAQAAALZYRcopVDAAAAAAGIYEAwAAAIBhaJECAAAAbLGKlFOoYAAAAAAwDAkGAAAAAMPQIgUAAADYYhUpp1DBAAAAAGCYdFnBeOzJqq4OAXdxeXIjV4eAFBRsv8jVISAF0X9ecnUIuIsyOfO7OgSk4PDV31wdAvCfky4TDAAAAOCB0SLlFFqkAAAAABiGBAMAAACAYWiRAgAAAGxZLK6OIE2jggEAAADAMCQYAAAAAAxDixQAAABgi1WknEIFAwAAAIBhSDAAAAAAGIYWKQAAAMAWLVJOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYMtCi5QzqGAAAAAAMAwJBgAAAADD0CIFAAAA2GIVKadQwQAAAABgGBIMAAAAAIahRQoAAACwZbG4OoI0jQoGAAAAAMOQYAAAAAAwDC1SAAAAgC1WkXIKFQwAAAAAhiHBAAAAAGAYWqQAAAAAW7RIOYUKBgAAAADDkGAAAAAAMAwtUgAAAIAtCy1SzqCCAQAAAMAwJBgAAAAADEOLFAAAAGDDYra4OoQ0jQoGAAAAAMOQYAAAAAAwDC1SAAAAgC0etOcUKhgAAAAADEOCAQAAAMAwtEgBAAAAtnjQnlOoYAAAAAAwDAkGAAAAAMPQIgUAAADY4kF7TqGCAQAAAMAwJBgAAAAADEOLFAAAAGCLB+05hQoGAAAAAMOQYAAAAAAwDC1SAAAAgC1apJxCBQMAAACAYUgwAAAAABiGFikAAADAloUH7TmDCgYAAAAAw5BgAAAAADAMLVIAAACALVaRcgoVDAAAAACGIcEAAAAAYBhapAAAAABbZlaRcgYVDAAAAACGIcEAAAAAYBgSjEesZYvGOnx4i65cPqqNGxarbNlSd51fr15NRUWu1ZXLRxW+a5VqVK9s9/M6dWpo6S+zdO73KCX8fVYlShR9iNGnf3N3HtOr3yxT+cE/6v0pYdr7+8UU5zabvk6lBi5M9mo7Z5PD+YOWRqjUwIWauf3owwo/XWvc7G1t3bNSx86Fa8mq2SpVpvhd59esU03rti3WsXPhWr3pR1Wp+oLdz3+7uM/hq2W7Jg/zMNKtVi0b69iRbfrz6nFt2bRE5e5xbatf/3Xt27tef149rt0Rq/VqjSrJ5vTr20VnT0fojyvHtHL5XBUo8NRDij59q9+4rn7cNkfrjq/U5CXjVLRUkRTnPlUovwZ/218/bpujrb+vVcOP6t913x+0eUdbf1+rjv3bGB32f0Kz5u9pz761Ohe3T6vWLFSZ4BJ3nV+nbg1tC1+hc3H7tGnbL6parVKKc78eOUAX/ziqlq0/NDjq/wiL2X1faQAJxiP05pu1NHRob33++UiFhL6mvXsP6JclMxQQkNPh/NDQYM2YPkbffTdXISGvavGSlVqwYLKKFi1snePr66PNW3aoZ6/Bj+ow0q2V+8/q61VRavFiUc1pXlWFgrKr9eyNunjtb4fzhzd4Tqs7vW59LWzxijKYTHql6OPJ5q459Luifr+ggCyZHvZhpEu13qihPoM+1Yih4/Vq5QY6sO+wZi6cqJz+fg7nB5cvpbGThmrurJ9U46UGWrFsjSbP/EaFnylgnVO6SCW7V+e2vWQ2m7Vs8apHdVjpRoMGtTXsq74aOGi4yoXUUGTUAS1bOivFa1uF0LKaNWOspk2bo7Llq2vx4pX6YeEUFSt2+9rWtUtrtW3TVK3bdtNzFWvp2vXrWvbLLHl7ez+qw0oXXq5dWe37ttKU4d/rwxof6+iB4xoxa6hy5MzucH6mzN46d+acxg3+VvExF+6672dKFlbd92vp6IHjDyHy9O+Neq9p0JAeGvrFGFWuWFf79h3Uwp+myj+F61r5kNKaNG2EZk1fqJcq1tGyX1Zr5pxxeuaZgsnm1qz1isqWK6Vz56If9mEADpFgPEId2jfX1KlzNH36fB06dFRt2nbX9et/q3Hjhg7nt23TTL/+uk7DR0zUocPH1L//MO3evU+tWzW2zpk9+0cNHjxKa9Y4/tYc92/GtiOqV/op1S2VX/8LyKpeNcsok1cG/bznlMP52TJnlP9jmayvbSdjlckrg6o9Y59gxFz9S1+s2KPBdcvL04O33IP4uHUjzZm+UPNn/6yjh0+oW+cB+vv633r7vTcczm/W4n2tC9usCaOn6diRExo2eIz2RR3Qhx+9a50TF3vB7lXt1crasnGHzpz+7VEdVrrRqUNzTZ4yW99Pn6+DB4+qdZtuun79LzX58G2H89u1a6aVK9fp6+ETdOjQMfXt99U/17bb1aP27T7S4CGjtGTJr9q796A+bNJBefIEqU6d6o/qsNKFd5o30OLZS7V0/gqdOnpaQ7sNV8Jff+v1t191OP9g5GGNGTRRqxevVeKNxBT3m9knk/qN6akvPh2mPy7/8bDCT9dat22q6d/N0+yZP+jw4WPq3KGPrv/1l95r9KbD+S1aNVbY6o0aPWqyjhw+rsGDRioq8oA+avGB3bzcuYP05Vd91KJZZyUlJj2KQwGS4dPOI+Ll5aUyZZ61SwQsFovWrN2o0JBgh9uEhJZJljisWr1eISnMx4NLvGnWwfOXFfJUoHXMw2RSyFNBivrt7t/i/evn3SdVvVg+Zc54e3E2s8WiXot2qHGFQioQmM3wuP8LvLw89WzJotq4fpt1zGKxaOP6bSpTrqTDbYLLldTG9Vvtxtav2aLgFOb7B+TUy9Ve1NyZPxoX+H/ErWtbCYWt2Wgds1gsCluzSaGhjq9VoSHBdvMl6ddV66zzn3rqCeXOHaQwm+vf1at/aMeO3SleL5Gcp5enCpcopJ0bw61jFotFOzdFqHhwMaf23WVwR20J26adGyOcDfM/ycvLSyVLF9P6dVusYxaLRevXbVG58qUdblOufGmtX7vFbmzN6o0qV76U9c8mk0njJ32l0aMm69ChYw8l9v8Ms8V9X2kACcYj4u/vJ09PT8XExtmNx8bEKygowOE2uYICFBMbf9/z8eAuXU/QTYtFOR+zb2HK6eut+D8dt0jZ2vv7RR2Lu6o3Stv3iE/bfFgZPEx6t3yBFLbEvfjlzCFPT0/FxdknevFxFxQY5O9wm4BAf8XH2s+Pi41XQKDj+Q3erq1rf17X8l9WGxP0f8i/17bYmDuuVbFxypXStS1XQLJrYUxMvHV+rqDAf8bumBMbr1y5AoX7k90vmzw9M+hi/CW78Ytxl5QzwHEbzv2oWruyChcvqPFDJjkb4n9Wzn+va3f8Gx8Xe0FBgY7fN4FB/oq98zNBbLwCbd5nHTp/rJtJNzVx/PfGBw2kglsnGGfPnlXTpk3vOichIUFXr161e1ksaSO7Q/rx856TKhiYTc/mvf2P9oHzlzR7x1ENqF1OJpPJhdHhXhq+94Z+WvCLEhJuuDoUwK0F5glQpwFt1bfd57qRkHILFR69kqWKqUWrxmrT8jNXhwK494P2Ll68qO+//15Tp05Ncc6QIUPUv39/uzGPDFnk6ele7Sjx8ReVlJSU7JuJwCD/ZN/S/Ss6Jk5Bd3zjerf5eHA5fLyVwWTShTuqFReuJcj/sbvfmP3XjSSt3H9WrSrZtxxEnInXxWsJenXUMuvYTYtFw1dFatb2o1re/jXjDiAdu3jhkpKSkpLdMOwfkDPZt+b/iouNl3+g/fyAQP9k3xZKUvnQMipQ6Gm1atbVuKD/Q/69tt1ZTQoMDFB0Ste26Lhk18KgIH/r/OiY2H/GAhQdHXt7TqC/9kTuNzL8dO3yxStKSropP/8cduN+ATl0IS7lFfLupsizheQX4KfvVnxrHfP0zKBSoSVU/8M3VOmpajKb08YqN6504d/r2h3/xgcE5kxW3ftXbEy8Au/8TBDor9h/3jcVniungICcijq43vpzT09PDRzcTS1bN1ap4varUOLuLPx37BSXJhiLFy++689PnDhxz310795dnTt3thvzD3C/pVoTExMVEbFXlSs/r8VLVkq61StZ+aWKGj/hO4fbbN8WocqVn9foMVOsYy9XeUHbt4c7nI8H55XBQ8/kzq4dp2JVpUheSbfun9hxMlZvl/vfXbf99eBvupFkVs1nn7Abf/3ZJxT6lH07R6vZG/X6s0+qTsn8hsafniUmJmlv5AFVfDFEK5etkXTrvVOxUoi+mzTH4TbhOyNV8cVQTZkw0zr2wksVFL4zMtnct9+vp8jd+3Vw/+GHcwDp3K1rW5SqVK6oxYtvX9uqVK6oceOnOdxm2/ZwValSUd+Mnmwdq/ryi9q27da17eTJMzp/PkZVKldU5D8JRZYsj6l8+dKa8O30h3xE6UdSYpIORx1R2YpltGHlZkm3zk3ZimW0cNpPD7TPXZsi9F4V+6Wcew7/TKePn9HMsXNILu5TYmKiInfv14uVKmjZP62ZJpNJlSo9p0nfznC4zc4du/XiSxU0Ydx31rGXqjyvnTv2SJLmzf1Z69dutttmwc9TNX/uIs2e+cNDOQ4gJS5NMOrWrSuTyXTXlqZ7tZZ4e3snW7bQXdtRRn0zSVMmD1d4RJR27dyjdu2aydc3s6ZPny9JmjJlhM6di1bv3l9KksaMnaLVqxaoY4ePtXx5mBq8VVvBwSXUuk036z5z5MiufPnyKE/uIElSoUK3PgzHxMRR6UilD0ILqfeinSqaO4eK5/HTrB1H9VdikjUZ6PXzDgVmyaz2Lz9rt93Pu0+qcuE8yu5j/99hdh/vZGOeHh7K+Vgm5ffP8lCPJb35dtx0jRj7uSL37NeeiH36qOX7yuyTWfNm/yxJGjlusKLPx+qLgSMlSVMmztTCJdP0cZvGCvt1g+rUe1UlShXTZ5362e33sSy+er1ONQ3oPezRHlA6M2LUJE2bMkLhEVHauXO32rdrLl/fzPru+3mSpGlTR+ncufPq2esLSdLo0VO0JmyhOnVsoWXLV6vhW3UUHFxCLVt/at3nN6Mnq0f39jp67IROnTqr/v266ty5GC1atNIlx5hWzZm0QL1HdNOhqCPav/ug3m7+pjJlzqRf5q2QJPUZ1V1x5+M0/otbyZ6nl6eeKvSk9f8H5PJXwWL/01/X/tJvp87p+rW/dOLwKbvf8ff1v3X10tVk47i7cWOmauzEodqze58iwqPUsvWH8vHJrNkzbiUD4yYO1fnzMRrY72tJ0sTx32vJ8llq066pfl25TvXq11Sp0sXVqV0vSdKli5d16eJlu9+RlJik2Jh4HTt68pEeG+DSBCN37twaN26c6tSp4/Dne/bsUXBw+lkxZOHCJQrw91OfPp8oV1CAIiMPqFbtD6w3beXLl1dmm9UBtm0LV6PG7dS/X1cNGPCpjh07pQYNPtKBA7e/aX399Vc0edJw659nzRwnSRo4aLgGDRrxiI4sfaheLJ8uXU/Q+PUHFP/n3yoclE3j3q1ovfH7/NXryZLXU/F/aPfZCxr/3guOdgmDLPlphXLmzKEu3dsqINBfB/Yd0gcNWir+nxu/8z6e2+6b0/Ade9T248/0aY92+qxXB508cVofvd9ehw/ar6pSp96rMplMWvTDMuHBLViwWAH+furXp4ty5QpQZOR+1Xz9feu17Yl8eezOz9Ztu/R+o7Ya0P9TDRr4mY4eO6n6bzbTfpsq0lfDxsnX10cTxg1V9uxZtXnzTtWs9b4SEhIe+fGlZWGL1yqHXzZ91OVD5Qzw09H9x9Xp/c906Z8bv4PyBNqdG/+gnJr+6+3K0nut3tZ7rd5WxJY9atOg0yOPPz376cdlyunvp+49OygwKED7og6qQb1m1gUtHs+XR2abL2B3bN+tj5t2Vo8+ndSr7yc6cfyU3n+ntQ4e5OGtD0UaWa3JXZksLrwjunbt2ipVqpQGDBjg8OeRkZEqXbp0qkuu3pnyGREeHpLLkxu5OgSkoGD7Ra4OASmI/vPSvSfBZcoFFHJ1CEjB4as828ZdXfzDfZOja5+772cV357u3yrq0gpG165dde3atRR/XqBAAa1du/YRRgQAAADAGS5NMF544e5tJb6+vqpUqdIjigYAAACQZGHBAme49XMwAAAAAKQtJBgAAAAADOPWD9oDAAAAHjlWkXIKFQwAAAAAhiHBAAAAAGAYWqQAAAAAW6l8BhvsUcEAAAAAYBgSDAAAAACGoUUKAAAAsMUqUk6hggEAAADAMCQYAAAAAAxDixQAAABgy8IqUs6gggEAAADAMCQYAAAAAAxDixQAAABgi1WknEIFAwAAAIBhSDAAAAAAGIYWKQAAAMCGxcwqUs6gggEAAADAMCQYAAAAAAxDixQAAABgi1WknEIFAwAAAIBhSDAAAAAAGIYWKQAAAMAWLVJOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYMvCg/acQQUDAAAAgGFIMAAAAAAYhhYpAAAAwBarSDmFCgYAAAAAw5BgAAAAADAMLVIAAACADQstUk6hggEAAADAMCQYAAAAAAxDixQAAABgixYpp1DBAAAAAGAYEgwAAAAAhqFFCgAAALBlNrs6gjSNCgYAAAAAw5BgAAAAADAMLVIAAACALVaRcgoVDAAAAACGIcEAAAAAYBhapAAAAABbtEg5hQoGAAAAAMOQYAAAAAAwDC1SAAAAgA2LhRYpZ1DBAAAAAGAYEgwAAAAAhqFFCgAAALDFKlJOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYIsWKadQwQAAAABgGBIMAAAAAIahRQoAAACwYaFFyinpMsG4aTa7OgTcRenOv7o6BKTg2NKerg4BKcj9Sm9Xh4C7OPbnOVeHgBRcTbju6hCA/xxapAAAAAAYJl1WMAAAAIAHRouUU6hgAAAAAOnU2LFjlT9/fmXKlEkhISHasWPHXedfvnxZbdq0Ue7cueXt7a1ChQpp2bJlqfqdVDAAAACAdGjevHnq3LmzJkyYoJCQEI0cOVLVq1fX4cOHFRgYmGz+jRs39MorrygwMFALFy5U3rx5dfr0aWXPnj1Vv5cEAwAAALCVTtYLGj58uJo3b64mTZpIkiZMmKClS5dq6tSp6tatW7L5U6dO1cWLF7VlyxZ5eXlJkvLnz5/q30uLFAAAAJDO3LhxQ+Hh4apatap1zMPDQ1WrVtXWrVsdbrN48WJVqFBBbdq0UVBQkIoXL67Bgwfr5s2bqfrdVDAAAACANCIhIUEJCQl2Y97e3vL29rYbi4+P182bNxUUFGQ3HhQUpEOHDjnc94kTJ7RmzRq99957WrZsmY4dO6bWrVsrMTFRffv2ve8YqWAAAAAANixmi9u+hgwZomzZstm9hgwZYshxm81mBQYG6ttvv1VwcLAaNmyonj17asKECanaDxUMAAAAII3o3r27OnfubDd2Z/VCkvz9/ZUhQwbFxMTYjcfExChXrlwO9507d255eXkpQ4YM1rFnnnlG0dHRunHjhjJmzHhfMVLBAAAAANIIb29vZc2a1e7lKMHImDGjgoODFRYWZh0zm80KCwtThQoVHO77+eef17Fjx2Q2377L/ciRI8qdO/d9JxcSCQYAAABgz2xx31cqdO7cWZMmTdL333+vgwcPqlWrVrp27Zp1ValGjRqpe/fu1vmtWrXSxYsX1aFDBx05ckRLly7V4MGD1aZNm1T9XlqkAAAAgHSoYcOGiouLU58+fRQdHa1SpUppxYoV1hu/z5w5Iw+P2/WGfPnyaeXKlerUqZNKlCihvHnzqkOHDvrss89S9XtNFosl3T0L3TNjXleHgLsokD2Pq0NACvYs6nzvSXCJ3K/0dnUIuAtPm35luJdLf/3p6hCQgqQbv7s6hBRdfqeyq0NIUfY5a10dwj1RwQAAAABspZMH7bkK92AAAAAAMAwJBgAAAADD0CIFAAAA2LCkcrUm2KOCAQAAAMAwJBgAAAAADEOLFAAAAGCLVaScQgUDAAAAgGFIMAAAAAAYhhYpAAAAwAarSDmHCgYAAAAAw5BgAAAAADAMLVIAAACALVaRcgoVDAAAAACGIcEAAAAAYBhapAAAAAAbFlqknEIFAwAAAIBhSDAAAAAAGIYWKQAAAMAWLVJOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYINVpJxDBQMAAACAYUgwAAAAABiGFikAAADAFi1STqGCAQAAAMAwJBgAAAAADEOLFAAAAGCDVaScQwUDAAAAgGFIMAAAAAAYhgQDAAAAgGG4BwMAAACwwT0YzqGCAQAAAMAwJBgAAAAADEOLFAAAAGCDFinnUMEAAAAAYBgSDAAAAACGoUUKAAAAsGUxuTqCNI0KBgAAAADDkGAAAAAAMAwtUgAAAIANVpFyDhUMAAAAAIYhwQAAAABgGFqkAAAAABsWM6tIOYMKBgAAAADDkGC4QKuWjXXsyDb9efW4tmxaonJlS911fv36r2vf3vX68+px7Y5YrVdrVEk2p1/fLjp7OkJ/XDmmlcvnqkCBpx5S9Onbu00bKGzXIkWe2aR5y6fp2dJFU5xboPDT+mbqlwrbtUiHYneq0cfvOL1PpGzu6h169ZMRKvfRQL03YJL2nvjtrvNnrtyq2t1Gq3zzQarWebi+mr1CCTcSrT+f8stGvdv/W1VoOVgvtRuqjqPm6NT5+Id9GOnWRx+/r8j963Q+fr9WrV2oMsEl7jq/zhuvanvESp2P36/N25fqlWqVUpw7fNQAXfrzmFq2/tDgqP8bmn70rsKjwnQ2JkorwuardJln7zq/dt0a2rJzuc7GRGn9lsWq+sqLdj8fPW6I4q4ctnvN+2HywzyEdIvPA0ivSDAesQYNamvYV301cNBwlQupocioA1q2dJYCAnI6nF8htKxmzRiradPmqGz56lq8eKV+WDhFxYoVts7p2qW12rZpqtZtu+m5irV07fp1Lftllry9vR/VYaULr9Z5Rd36d9TYYZNVr+oHOrz/qCbPGy0//xwO52fKnElnT/+urweNUWyM4w+mqd0nHFuxfZ+GzV2pFnVf0tz+LVQ4X5BaDZupC1f/dDh/2dYojVqwWi3rVNJPg9uoX9PaWrljn775Icw6Z9ehU2pYpZxm9P5IE7s2UtJNs1oOm6HrCTce1WGlG2/Uf02DhvTQl0NG66WKdbRv3yH98PM0+Qf4OZxfPqS0Jk8boZnfL1Cl52tr6S+rNHPueD1TtGCyuTVrvaKy5Urp3Lnoh30Y6VLdeq9qwODuGvblWL384hvav++Q5v80Rf7+js9NufKlNXHK15o1Y6GqvFBXy5eG6fvZY1XkGftzE7Zqg4oVfN76+rhZ50dxOOkKnwfcm8Xsvq+0wGSxWCyuDsJonhnzujqEFG3ZtEQ7d0WqQ8dekiSTyaRTJ3Zq7LhpGvrV2GTzZ88aL18fH9V5o7F1bPPGJdoTuV9t2naTJJ09HaERIydq+IiJkqSsWbPo3G971PSjTpo/f/EjOKrUKZA9j6tDcGje8mnat+eABnb/StKtc7Nuzy+aOXm+Jo3+/q7bhu1apO+/navp384xbJ+usGeRe35IeG/AJBV7Ko96fFBTkmQ2m1Wt8wi9U7W8mr3+QrL5g2cs1clz8Zr02e33zbA5K7X3xG/6vmczh7/j4tVrqtz+K03t/qGCC+d/KMfhjNyv9HZ1CClatXahdkfs1aef9Jd067/zfYc3atKEGRo5fGKy+VO+HyVfn8x6u8HH1rFf1yzUvr0H1LlDH+tY7txBWrXuB71Zt4nmLZyk8WO/04Rx3z3043kQnhkyuDoEh1aEzdeeiL3q1nWgpFvnJvLAek3+doa+GTEp2fxJ00bIxyez3mvY0jq2fPU87dt7SF079ZV0q4KRNVtWNX6vzaM5CCdd+svxFxGuxucBKenG764OIUXnnqvs6hBSlGfLWleHcE9UMB4hLy8vlSlTQmFrNlrHLBaLwtZsUmhosMNtQkOC7eZL0q+r1lnnP/XUE8qdO0hhazZZf3716h/asWO3QkMc7xPJeXl5qljJItqyYYd1zGKxaOuGHSpV9u7tBI9yn/9FiUlJOnjqnEKLPm0d8/DwUGixpxV13HGbVKkC+XTw1DlrG9VvsRe1KeqoXiiR/Bvyf/3519+SpKy+mQ2MPv3z8vJSqdLFtW7tZuuYxWLR+rVbVK58aYfblC9fWuvWbrEbWxO20W6+yWTShMnDNHrUJB06ePThBJ/OeXl5qWSpYlq/7vbftcVi0YZ1W1S2nONzU7ZcKW1Yt9VubG3YJpUtV8pu7PmK5XXg2BZt3bVCQ4f3U44c2Y0OP13j8wDSO5evIvXXX38pPDxcfn5+KlrUvjf977//1vz589WoUSMXRWcsf38/eXp6JmuniY2NU5HC/3O4Ta5cAYqJjbMbi4mJV66ggFs/Dwr8Z+yOObHxypUr0KjQ070cftnl6empC3EX7cbj4y7qqQL53Waf/0WX/rium2aLcmZ7zG48Z1ZfnUzhnonXKpTQpT+v68PPp0qSkm6a1aByWX1U60WH881ms4bOXqFSBfOp4ONBxh5AOpczZw55enoqLvaC3XhcbLwKFnra4TaBQf6Ki4tPNj/wn+uaJHXs3EJJSTc1cZz7VfrSCr8Uzk1s3AUVuMu5iY2949zEXVBgkL/1z2FhG/XLklU6c/o35X8qn3r26ay5P0zSq1UbymxOI/0bLsbnAfdnsbCKlDNcmmAcOXJE1apV05kzZ2QymVSxYkXNnTtXuXPnliRduXJFTZo0uWuCkZCQoISEBLsxi8Uik4n/MID/qp0HT2rKko3q2aimnn36cZ2Jvaihs5Zr4qL1alEn+c3Eg2cs0/HfYvVdz6YuiBZ3KlmqmFq0bqyXnq/j6lDgwM8/LLP+/4MHjujA/sPaFRmm518or43rt7kwMgDuwqUtUp999pmKFy+u2NhYHT58WFmyZNHzzz+vM2fO3Pc+hgwZomzZstm9LOY/HmLUDy4+/qKSkpLsvgmSpMDAAEXf8Y3Dv6Kj4xQUGGA3FhTkb50fHRP7z9gdcwL9FR0da1To6d6li5eVlJSknHfclOof4Kf4O779c+U+/4tyZPFRBg+TLlyx76O+cPWa/O+oavxr7E9r9fpzJVWvUrAK5gvSy8HPqN2bL2vq0o3JvmEdPGOpNkQe0aRuHyrIL9tDO4706sKFS0pKSlJAoP2NqQGB/ikufhAbE6+AAH8H829d1yo8V04BATm199AGxV0+pLjLh/TEk49r0JDuity/7qEcR3p0MYVzExiQ867nJjDwjnNzl/mSdPrUb4qPv6innn7S+aD/I/g8gPTOpQnGli1bNGTIEPn7+6tAgQJasmSJqlevrhdeeEEnTpy4r310795dV65csXuZPLI85MgfTGJioiIiolSlckXrmMlkUpXKFbVtW7jDbbZtD1eVKhXtxqq+/KJ1/smTZ3T+fIzdPrNkeUzly5fWtu2O94nkEhOTtD/ykCq8UM46ZjKZFPpCOe3Ztddt9vlf5OXpqWfy59H2AyetY2azWdsPnFCJ/z3ucJu/ExJl8rCvYmbwuHW5+3dVC4vFosEzlmpN+CFN+rSxHg9gZa8HkZiYqD2796nSS89Zx0wmk1586Tnt3LHb4TY7duy2my9JlSs/b50/b+7PqhhaUy8+V8v6OncuWqNHTlb9uk0e3sGkM4mJiYrcs18vVqpgHTOZTHqhUgXt2un43OzauUcvVAq1G6tU+Tnt2rknxd+TO0+Q/PyyKyba8QdjJMfnAffn6pWi0voqUi5tkfrrr7/k6Xk7BJPJpPHjx6tt27aqVKmSZs+efc99eHt7J1t+zZ3bo0aMmqRpU0YoPCJKO3fuVvt2zeXrm1nffT9PkjRt6iidO3dePXt9IUkaPXqK1oQtVKeOLbRs+Wo1fKuOgoNLqGXrT637/Gb0ZPXo3l5Hj53QqVNn1b9fV507F6NFi1a65BjTqu8mzNYXo/tqX+RBRUXsV+MW7yizT2b9OHeJJOmLMf0Uez5Owz+/tbqHl5en/lf4Vh+zV0YvBeUOUJHihXT92nWdOfnbfe0T9+eD6hXUe9JPKvZUHhV/Oq9m/rpNfyUkqu4Lt25U7fntjwrMkVUdGlSVJFUqVUgzVm5VkSdy6dn/Pa6zMRc19sc1erFUYWuiMXjGUi3fulcjO7wj30wZFX/5VuXzMZ9MypTRyzUHmkaNGzNV4yZ+pd0RexURHqVWbT6Ur09mzZq5UJI0/tuvdP5cjAb0GyZJmjjuO/2yYrbatGumX1euVb03X1epMsXVsX1PSbeqf5cuXrb7HUmJSYqJidOxoyeF+zdh7DSNHv+l9uzep4jwKLVo3Vg+vpk1Z+aPkqQxE75U9PkYDeo/XJL07fjpWrRshlq1baJVK9frjfqvqVTp4vrkn9W9fH191KVbW/2yaKViY+OV/6l86jugq06eOK21YRtTjAPJ8XkA6ZlLE4wiRYpo165deuaZZ+zGx4wZI0mqXbu2K8J6qBYsWKwAfz/169NFuXIFKDJyv2q+/r71pron8uWxa+HYum2X3m/UVgP6f6pBAz/T0WMnVf/NZtq//7B1zlfDxsnX10cTxg1V9uxZtXnzTtWs9X6ye1Nwd8sXrZJfzuxq92kLBQTm1MF9R9T87fbWm7Tz5M0li/n2qs6BuQL085pZ1j83a/OBmrX5QDs2h6vRGy3va5+4PzVCiuvSH9c07qe1ir/ypwo/kUvjPnnfeuN39IUr8rD5YqF57RdlMpk09sc1ir30h3Jk8VGlUoXVtv7th1LNX7NLktTsi+/sfteAZnVU5wXHK+zAsZ9+WCZ//5zq0aujAoMCtDfqgN58o6n15uLH77iu7di+W82bdlbP3p3Uu98nOnH8lN5/u5UOHmC1KKP9/ONy5czpp896tFdgUID27T2ohvU+UlzcP+fm8dyy2JybnTt2q+VHXdS9V0f17NNZJ46fUuN321hX8rp586aKFSukhu/UVbZsWRR9Plbr1m7WF4NG6YbNgyxxb3weQHrm0udgDBkyRBs3btSyZcsc/rx169aaMGFCqlelcOfnYMB9n4MB930OBtz7ORhw3+dgwH2fgwH3fg7G2XIvuzqEFOXbGXbvSS7m0nswunfvnmJyIUnjxo1jyTsAAAAgDeFBewAAAAAM4/IH7QEAAADuxHU3EKQPVDAAAAAAGIYEAwAAAIBhaJECAAAAbFjM7vtMtbSACgYAAAAAw5BgAAAAADAMLVIAAACADVqknEMFAwAAAIBhSDAAAAAAGIYWKQAAAMAGD9pzDhUMAAAAAIYhwQAAAABgGFqkAAAAABusIuUcKhgAAAAADEOCAQAAAMAwtEgBAAAANiwWWqScQQUDAAAAgGFIMAAAAAAYhhYpAAAAwIbF7OoI0jYqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2zKwi5RQqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2eNCec6hgAAAAADAMCQYAAAAAw9AiBQAAANiwmGmRcgYVDAAAAACGIcEAAAAAYBhapAAAAAAbFourI0jbqGAAAAAAMAwJBgAAAADD0CIFAAAA2GAVKedQwQAAAABgGBIMAAAAAIahRQoAAACwYbbQIuUMKhgAAAAADHNfFYzFixff9w5r1679wMEAAAAASNvuK8GoW7fufe3MZDLp5s2bzsQDAAAAuJSFFimn3FeCYTabH3YcAAAAANIB7sEAAAAAYJgHWkXq2rVrWr9+vc6cOaMbN27Y/ax9+/aGBAYAAAC4gsXi6gjStlQnGLt379Zrr72m69ev69q1a/Lz81N8fLx8fHwUGBhIggEAAAD8h6W6RapTp06qVauWLl26pMyZM2vbtm06ffq0goODNWzYsIcRIwAAAIA0ItUVjD179mjixIny8PBQhgwZlJCQoKefflpDhw5V48aNVa9evYcRJwAAAPBI8KA956S6guHl5SUPj1ubBQYG6syZM5KkbNmy6ezZs8ZGBwAAACBNSXUFo3Tp0tq5c6cKFiyoSpUqqU+fPoqPj9eMGTNUvHjxhxEjAAAAgDQi1RWMwYMHK3fu3JKkzz//XDly5FCrVq0UFxenb7/91vAAAQAAgEfJYjG57SstSHUFo2zZstb/HxgYqBUrVhgaEAAAAIC0iwftAQAAADBMqisYTz31lEymlMszJ06ccCogAAAAwJV40J5zUp1gdOzY0e7PiYmJ2r17t1asWKGuXbsaFRcAAACANCjVCUaHDh0cjo8dO1a7du1yOiAAAAAAaZdh92C8+uqr+uGHH4zaHQAAAOASZovJbV9pgWEJxsKFC+Xn52fU7gAAAACkQQ/0oD3bm7wtFouio6MVFxencePGGRocAAAAgLQl1QlGnTp17BIMDw8PBQQE6KWXXlKRIkUMDQ7p06mrMa4OASl4rFIXV4eAFPx1bqOrQ8BdFChc19UhIAWX9KerQ0AalFYeaOeuUp1g9OvX7yGEAQAAACA9SPU9GBkyZFBsbGyy8QsXLihDhgyGBAUAAAAgbUp1BcOSwpNHEhISlDFjRqcDAgAAAFwprazW5K7uO8H45ptvJEkmk0mTJ0/WY489Zv3ZzZs3tWHDBu7BAAAAAP7j7jvBGDFihKRbFYwJEybYtUNlzJhR+fPn14QJE4yPEAAAAECacd8JxsmTJyVJlStX1o8//qgcOXI8tKAAAAAAV3F8QwDuV6rvwVi7du3DiAMAAABAOpDqVaTq16+vL7/8Mtn40KFD1aBBA0OCAgAAAJA2pTrB2LBhg1577bVk46+++qo2bNhgSFAAAACAq5gtJrd9pQWpTjD+/PNPh8vRenl56erVq4YEBQAAACBtSnWC8eyzz2revHnJxufOnauiRYsaEhQAAACAtCnVN3n37t1b9erV0/Hjx1WlShVJUlhYmGbPnq2FCxcaHiAAAADwKFnSSCuSu0p1glGrVi39/PPPGjx4sBYuXKjMmTOrZMmSWrNmjfz8/B5GjAAAAADSiFQnGJJUs2ZN1axZU5J09epVzZkzR126dFF4eLhu3rxpaIAAAAAA0o5U34Pxrw0bNqhx48bKkyePvv76a1WpUkXbtm0zMjYAAADgkTO78SstSFUFIzo6Wt99952mTJmiq1ev6q233lJCQoJ+/vlnbvAGAAAAcP8VjFq1aqlw4cKKiorSyJEjde7cOY0ePfphxgYAAAAgjbnvCsby5cvVvn17tWrVSgULFnyYMQEAAAAuYxGrSDnjvisYmzZt0h9//KHg4GCFhIRozJgxio+Pf5ixAQAAAEhj7jvBCA0N1aRJk3T+/Hm1aNFCc+fOVZ48eWQ2m7Vq1Sr98ccfDzNOAAAAAGlAqleR8vX1VdOmTbVp0ybt3btXn3zyib744gsFBgaqdu3aDyNGAAAA4JExW9z3lRY88DK1klS4cGENHTpUv/32m+bMmWNUTAAAAADSKKcSjH9lyJBBdevW1eLFi43YHQAAAIA06oGe5A0AAACkV2ZWkXKKIRUMAAAAAJBIMAAAAAAYiBYpAAAAwAYP2nMOFQwAAAAAhiHBAAAAAGAYWqQAAAAAG2ZXB5DGUcEAAAAAYBgSDAAAAACGoUUKAAAAsMEqUs6hggEAAADAMCQYAAAAAAxDixQAAABgg1WknEMFAwAAAIBhSDAAAAAAGIYWKQAAAMAGLVLOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYIMH7TmHCgYAAACQTo0dO1b58+dXpkyZFBISoh07dtzXdnPnzpXJZFLdunVT/TtJMAAAAIB0aN68eercubP69u2riIgIlSxZUtWrV1dsbOxdtzt16pS6dOmiF1544YF+LwkGAAAAYMNsct9XagwfPlzNmzdXkyZNVLRoUU2YMEE+Pj6aOnVqitvcvHlT7733nvr376+nn376gf7+SDAAAACANCIhIUFXr161eyUkJCSbd+PGDYWHh6tq1arWMQ8PD1WtWlVbt25Ncf8DBgxQYGCgmjVr9sAxkmAAAAAAacSQIUOULVs2u9eQIUOSzYuPj9fNmzcVFBRkNx4UFKTo6GiH+960aZOmTJmiSZMmORUjq0gBAAAANsxuvIpU9+7d1blzZ7sxb29vp/f7xx9/6IMPPtCkSZPk7+/v1L5IMAAAAIA0wtvb+74SCn9/f2XIkEExMTF24zExMcqVK1ey+cePH9epU6dUq1Yt65jZfOuZ5p6enjp8+LD+97//3VeMtEgBAAAA6UzGjBkVHByssLAw65jZbFZYWJgqVKiQbH6RIkW0d+9e7dmzx/qqXbu2KleurD179ihfvnz3/bupYAAAAAA2LK4OwCCdO3dW48aNVbZsWZUvX14jR47UtWvX1KRJE0lSo0aNlDdvXg0ZMkSZMmVS8eLF7bbPnj27JCUbvxcSDAAAACAdatiwoeLi4tSnTx9FR0erVKlSWrFihfXG7zNnzsjDw/iGJpPFYkkvSZqVZ8a8rg4Bd+HpkcHVISAFSeabrg4BKfjr3EZXh4C7KFC4rqtDQAp+/+OCq0NACpJu/O7qEFL0c653XR1CiupGz3Z1CPdEBQMAAACwYXZ1AGkcN3kDAAAAMAwJBgAAAADD0CIFAAAA2DCb3PdBe2kBFQwXaNWysY4d2aY/rx7Xlk1LVK5sqbvOr1//de3bu15/Xj2u3RGr9WqNKsnm9OvbRWdPR+iPK8e0cvlcFSjw1EOKPn1r0aKRDh/erMuXj2jDhkUqW7bkXefXq1dTkZFrdPnyEe3a9auqV69s9/M6dWrol19m6vffI/X332dUokTRhxl+usb7xn3N+WGJqtVvrDKVa+ud5h2198DhFOcmJiVp/NRZqtGgicpUrq16jVtr07ZddnN27dmrNp/2VeXa76n4868qbMOWh30I6VqjZg21afdyHf59p37+dZZKlrn7cpOv1X5FYdsW6fDvO7Vy4w+qXLWi3c9PX4hy+GrR9sOHeBTpE9c1pFckGI9Ygwa1Neyrvho4aLjKhdRQZNQBLVs6SwEBOR3OrxBaVrNmjNW0aXNUtnx1LV68Uj8snKJixQpb53Tt0lpt2zRV67bd9FzFWrp2/bqW/TLLkMfG/5e8+WYtDR3aW59/PlKhoTW1d+9BLVkyM8VzExoarOnTR+u77+YpJOQ1LVmyUgsWTFLRooWsc3x9fbRly0716jXkUR1GusT7xn0tX71eQ0d/q1ZN39OCqaNVuMBTatG5ly5cuuxw/uhvv9eCRcvVo1MrLZo5UW/VfU0dug/UwSPHrHP++utvFS7wtHp+0voRHUX69Xrd6uo1sKtGfTVBr1dpqIP7DmvGggnK6e/ncH5wuZIaPelLzZ/5k2pWfku/Llujb2eMUqEiBaxzyj5T2e7VpV1vmc1mLVuy6lEdVrrAdQ3pGcvUPmJbNi3Rzl2R6tCxlyTJZDLp1ImdGjtumoZ+NTbZ/NmzxsvXx0d13mhsHdu8cYn2RO5Xm7bdJElnT0doxMiJGj5ioiQpa9YsOvfbHjX9qJPmz1/8CI4qddx1mdoNGxYpPDxSnTr1kXTr3Bw7tl3jx3+nYcPGJZs/Y8ZY+fr6qF69Jtax9et/VlTUAbVr18Nu7pNPPq7Dh7eofPkaioo68HAPxAnuukwt7xv3Xab2neYdVbxIIWsyYDabVfWNRnr3zdr66IO3ks2vXPs9fdz4bb1Tv5Z1rGOPQfL2zqgv+36abH7x51/VqCG99fKLzz28gzCAuy5T+/OvsxS1e5/6fHbrSw6TyaRte3/Vd5PmaPyoqcnmj5k8VD4+mdX03XbWsZ9WztSBvYfUs8sgh7/j2xkj9dhjvnr3jeYP5yCc5K7L1HJdc+9lahfkfs/VIaSowflZrg7hnqhgPEJeXl4qU6aEwtbc/qBgsVgUtmaTQkODHW4TGhJsN1+Sfl21zjr/qaeeUO7cQQpbs8n686tX/9COHbsVGuJ4n0ju1rl5Vmts/h4tFovWrt2kkJAyDrcJDS1jN1+SVq/ekOJ8PBjeN+4rMTFRBw4fVWi5UtYxDw8PhZYtpch9Bx1ucyMxURkzZrQb8/bOqN1R+x9mqP9JXl6eerbkM9q0fpt1zGKxaNP67SpTznH7Z5lyJbVp/Xa7sQ1rtqQ43z/AT1VeeUHzZv5kXOD/AVzXkN6RYDxC/v5+8vT0VGxMvN14bGyccgUFONwmV64AxcTG2Y3FxMRb5+cKCvxn7I45sfHKlSvQqNDTPeu5ibU/NzEx8QpK4dwEBQUoNtm5iUtxPh4M7xv3denyVd28aVZOvxx24zn9cij+4iWH2zwfEqzpc3/U6bO/y2w2a8uOCIWt36K4CxcfRcj/KTly5pCnp6fiY+2/wY+PvaCAQH+H2wQE+is+7o75cSnPr/92HV3787pW/LLamKD/I7iuIb1z+SpSBw8e1LZt21ShQgUVKVJEhw4d0qhRo5SQkKD3339fVaokv4HJVkJCghISEuzGLBaLTNz9DwBup1uHFur35Teq9e7HMpmkfHlyq27NV/TTL7+6OjQ8gLfeq6ufFy5VQsINV4cCGIoH7TnHpRWMFStWqFSpUurSpYtKly6tFStW6MUXX9SxY8d0+vRpVatWTWvWrLnrPoYMGaJs2bLZvSzmPx7REaROfPxFJSUlKTDI/pugwMAARd/xjcO/oqPjFBRo/21GUJC/dX50TOw/Y3fMCfRXdHSsUaGne9Zzc8e3dEFB/sm+DfpXTEycApOdm4AU5+PB8L5xXzmyZ1WGDB66cEe14sLFS/K/o6rxL78c2fXNF320c/VP+vWH77VkziT5ZM6kx/PkehQh/6dcunBJSUlJ8g+0v2nYPzCn4u6o1v4rLjZe/nfcZOwf4Hh+udAyKlDwKc2d8aNxQf9HcF1DeufSBGPAgAHq2rWrLly4oGnTpundd99V8+bNtWrVKoWFhalr16764osv7rqP7t2768qVK3Yvk0eWR3QEqZOYmKiIiChVqXx7yT+TyaQqlStq27Zwh9ts2x6uKlXslwis+vKL1vknT57R+fMxdvvMkuUxlS9fWtu2O94nkrt1bvaqcuXnrWMmk0kvvfS8tm+PcLjNtm0RdvMlqUqViinOx4PhfeO+vLy8VLRwQW3ftcc6ZjabtT18j0oWf+au23p7Z1RQgL+Sbt7UqnWbVfmFCg852v+exMQk7Y08qOdfDLGOmUwmPf9iiCJ2RjrcJmJnpN18SXrhpVCH8xu+/4ai9uzXwf1HjA38P4DrGtI7l7ZI7d+/X9OnT5ckvfXWW/rggw/05ptvWn/+3nvvadq0aXfdh7e3d7Ll19y5PWrEqEmaNmWEwiOitHPnbrVv11y+vpn13ffzJEnTpo7SuXPn1bPXrcRq9OgpWhO2UJ06ttCy5avV8K06Cg4uoZatb6+28s3oyerRvb2OHjuhU6fOqn+/rjp3LkaLFq10yTGmVd98M1mTJ3+tiIi92rlzj9q1ayZfXx9Nnz5fkjRlygidOxet3r2/lCSNHTtVq1bNV4cOzbV8+Rq99VZtBQeXUJs23az7zJEjm/Lly6vcuYMkSYUK/U/SreoHlY77x/vGfTVq+IZ6fv61ihUpqOJFC2vm/J/1198JqlvzFUlS94HDFOifU51a3VptLWr/IcXEXVCRgk8rNu6Cxk2dKYvFoqbv3b72X7/+l878ds7659/PxejQkePKljWLctNLniqTx03X12MHKWrPAUVG7FXTFu/LxyezFsz+WZI0fNznij4fo6EDv5EkTZs4S/OWTFXz1o20ZtUG1XrjVT1bqpi6dRpgt9/HsviqZu1qGtRn2KM+pHSD65p7M7vvR8k0weX3YPybDHh4eChTpkzKli2b9WdZsmTRlStXXBXaQ7FgwWIF+PupX58uypUrQJGR+1Xz9fetNxc/kS+PzObbnX9bt+3S+43aakD/TzVo4Gc6euyk6r/ZTPv3336Q1VfDxsnX10cTxg1V9uxZtXnzTtWs9X6ye1NwdwsXLpG/v5/69OmsoKAARUYeUO3aH1jPTb47zs22beFq3Li9+vXrogEDPtWxY6fUoEFzHThw+9u8119/RZMmDbf+eebMW0sPDho0QoMGjXhER5b28b5xX69WraRLl69ozOSZir94UUUK/k8Tvh5obZE6HxMrD5svfRJu3NDoSd/rt3PR8smcWS9UKKchvbsqa5bHrHP2HTqqpu0+s/556OhvJUl1Xq2qz3t98oiOLH345eeVyumfQ527tVZAoL8O7DusRm+1UnzcrZvq8+TNZffeCd8ZqfYfd1OXnu3UtVd7nTpxRh9/0EFHDh2z22+tN2rIZJIW/7D8kR5PesJ1DemZS5+DUbJkSX355ZeqUaOGJGnfvn0qUqSIPD1v5T0bN25U48aNdeLEiVTt152fgwH3fQ4G3Pc5GHDf52DgFnd9Dgbc9zkYcO/nYMzJ477PwXjnnPs/B8OlFYxWrVrp5s3bH2iKFy9u9/Ply5ffcxUpAAAAwEhm0SPlDJcmGC1btrzrzwcPHvyIIgEAAABgBB60BwAAAMAwLr/JGwAAAHAnLrtBOZ2gggEAAADAMCQYAAAAAAxDixQAAABggwftOYcKBgAAAADDkGAAAAAAMAwtUgAAAIANs6sDSOOoYAAAAAAwDAkGAAAAAMPQIgUAAADY4EF7zqGCAQAAAMAwJBgAAAAADEOLFAAAAGCDB+05hwoGAAAAAMOQYAAAAAAwDC1SAAAAgA0etOccKhgAAAAADEOCAQAAAMAwtEgBAAAANmiRcg4VDAAAAACGIcEAAAAAYBhapAAAAAAbFh605xQqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2WEXKOVQwAAAAABiGBAMAAACAYWiRAgAAAGzQIuUcKhgAAAAADEOCAQAAAMAwtEgBAAAANiyuDiCNo4IBAAAAwDAkGAAAAAAMQ4sUAAAAYMNscnUEaRsVDAAAAACGIcEAAAAAYBhapAAAAAAbPGjPOVQwAAAAABiGBAMAAACAYWiRAgAAAGzQIuUcKhgAAAAADEOCAQAAAMAwtEgBAAAANiyuDiCNo4IBAAAAwDAkGAAAAAAMQ4sUAAAAYMNscnUEaRsVDAAAAACGIcEAAAAAYBhapAAAAAAbPGjPOVQwAAAAABiGBAMAAACAYWiRAgAAAGzwoD3nUMEAAAAAYBgSDAAAAACGoUUKAAAAsGGmScopVDAAAAAAGCZdVjB4urt7e9Yvv6tDQAqiLp50dQhIQeY8L7g6BNzFnzsnuToEpCBLueauDgH4z0mXCQYAAADwoHjQnnNokQIAAABgGBIMAAAAAIahRQoAAACwwRpSzqGCAQAAAMAwJBgAAAAADEOLFAAAAGCDVaScQwUDAAAAgGFIMAAAAAAYhhYpAAAAwIbZ5OoI0jYqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2zDxqzylUMAAAAAAYhgQDAAAAgGFokQIAAABs0CDlHCoYAAAAAAxDggEAAADAMLRIAQAAADbMrg4gjaOCAQAAAMAwJBgAAAAADEOLFAAAAGCDB+05hwoGAAAAAMOQYAAAAAAwDC1SAAAAgA0apJxDBQMAAACAYUgwAAAAABiGFikAAADABg/acw4VDAAAAACGIcEAAAAAYBhapAAAAAAbPGjPOVQwAAAAABiGBAMAAACAYWiRAgAAAGzQIOUcKhgAAAAADEOCAQAAAMAwtEgBAAAANnjQnnOoYAAAAAAwDAkGAAAAAMPQIgUAAADYsLCOlFOoYAAAAAAwDAkGAAAAAMPQIgUAAADYYBUp51DBAAAAAGAYEgwAAAAAhqFFCgAAALBhZhUpp1DBAAAAAGAYEgwAAAAAhqFFCgAAALBBg5RzqGAAAAAAMAwJBgAAAADD0CIFAAAA2GAVKedQwQAAAABgGBIMAAAAAIahRQoAAACwYXZ1AGkcFQwAAAAAhiHBAAAAAGAYEoxHrFXLxjp6ZJv+uHpcmzctUbmype46v37917V373r9cfW4dkesVo0aVex+Xrfuq1q2dLaiz+9T4o3fVbJksYcYffrX4MM3tGj7PG06sUrTfpmgoqWeSXHu04Xy68tJA7Vo+zztPLdB73zUINmc+o3qaPbqaVp7eLnWHl6uKYvH6bnKIQ/zENKtli0a6/DhLbpy+ag2blissvd479SrV1NRkWt15fJRhe9apRrVK9v9vE6dGlr6yyyd+z1KCX+fVYkSRR9i9Olfq5aNdezINv159bi23Oe1bd/e9frzn2vbq3dc2ySpX98uOns6Qn9cOaaVy+eqQIGnHlL06dvcFZv0apsBKvdeV73XY4T2Hjt91/kzl65X7Q6DVf69T1WtVX999d1PSriRaP35+PkrVPKtTnavOh2HPOzDSJf4TOC+LG78v7SABOMRatCgtr76qq8GDRqu8iE1FBV1QEuXzlJAQE6H8yuEltXMGWM1bdoclStfXYsWr9QPC6eoWLHC1jm+vj7avGWHevT4/FEdRrr1Su0q6ti3jSYP/04fVP9IRw8c0+jZw5QjZ3aH8zNlzqTfz5zTmMETFR9zweGc2PNxGjN4ohrVaK7GrzbXrs0RGjZtsJ4ulP/hHUg69OabtTR0aG99/vlIhYS+pr17D+iXJTNSfO+EhgZrxvQx+u67uQoJeVWLl6zUggWTVbRo8vdOz16DH9VhpFsNGtTWsK/6auCg4SoXUkORUQe07B7Xtln/XNvKlq+uxQ6ubV27tFbbNk3Vum03PVexlq5dv65lv8ySt7f3ozqsdGHFlt0aNv1ntXizuuZ++YkKP5lHrT6fqAtX/nA4f9mmcI2a/YtaNqiun0Z0U7+WDbVy6x59M2ep3bz/5culsG/7W1/fDWj3KA4nXeEzAdIzk8ViSRupUCp4Zczr6hAc2rxpiXbtilSHjr0kSSaTSSdP7NTYcdP01Vdjk82fNWu8fH18VPeNxtaxTRuXKDJyv9q07WY398knH9exo9tVtlw1RUbuf7gH4qRS/v9zdQgOTftlgg5EHtJXPUdKunV+ftm1UPOn/ajvx8y667aLts/T3EkLNWfygnv+ntX7f9E3g8Zr8R3/YLuDqIsnXR2CQxs3LFZ4eKQ6duot6da5OX5sh8aNn6Zhw8Ylmz9zxjj5+mbWG/WaWMc2rF+kqKj9atuuh93cJ598XEcOb1W58tUVFXXg4R6IE26a3feWwy2blmjnHde2U/9c24Y6uLbN/ufaVsfm2rZ54xLtsbm2nT0doREjJ2r4iImSpKxZs+jcb3vU9KNOmj9/8SM4qtT5c+ckV4fg0Hs9RqjY/55Qj2b1JUlms1nVWg3QO69WVLO6VZPNHzzlB538PUaT+rS2jg2bvkh7j57W9wPbS7pVwVi7c6/mf9X10RyEk7KUa+7qEBziM4GUeON3V4eQoo/yv+nqEFI0+dRCV4dwT25XwUiH+Y4kycvLS2XKlFDYmo3WMYvFojVrNik0NNjhNqEhwVpjM1+Sfl21LsX5eHCeXp4qUqKQdmzcZR2zWCzasTFczwYbU2L28PDQK3WqKLNPJu3dtc+Qff4X3HrvPKs1azZZxywWi9as3ajQEMfvhZDQMnbzJWnV6vUKSWE+HlxK17awe1zbwu5ybXvqqSeUO3eQwmzO4dWrf2jHjt0pnnMkl5iUpIMnflPos4WsYx4eHgp9tqCijjhukypVOL8OnjhrbaP6LSZem3Yf0Aul7dtFT0fHq2qLvnqt7UB1/2aGzsdfengHkg7xmcD9md34lRa43TK13t7eioyM1DPPpNz7nhb5+/vJ09NTsTHxduMxsXEqXNjxN/q5cgUoJjbObiw2Jl5BQQEPLc7/qux+2eTp6amLcfb/SF6Mv6j8BZ5wat//K/K0pi4Zp4zeGfXXtb/UtVkvnTx69x5o3Pbve8fRe6FwoQIOt8kVFKCY2Phk83nvGC+la1tsbJyKpOLaFhMTr1z/nJ9cQYH/jN0xJzZeuXIFGhV6unfp6jXdNJuVM3sWu/Gc2bPo5LlYh9u8VjFYl65e04e9R0uyKOmmWQ1eeU4f1XvFOufZgk9qYOt3lD9PoOIuXdXEhSvVpM9o/fD1p/LNnOlhHlK6wWcCpHcuSzA6d+7scPzmzZv64osvlDPnrR7E4cOH33U/CQkJSkhIsBuzWCwymUzGBAo46fTxM3rvlWZ6LIuvXn79JfUb1UMt6rUjyQDgdnbuP6YpP61Wz4/e1LMFn9CZ6HgNnfaTJi78VS3erCZJqmhTzSj0ZB49W/BJvdp6gFZu3aN6VUJdFToAN+KyBGPkyJEqWbKksmfPbjdusVh08OBB+fr63leSMGTIEPXv399uzOTxmDJkyGpkuE6Lj7+opKQkBQb5240HBQYo+o5v6f4VHR2noED7byYCg/yTfasH512+eEVJSUnyC8hhN+7n76cLcRed2ndSYpJ+O3Wrz/TQ3iMqWqqI3v6ogYZ8Nsyp/f5X/PveSc17ITomTkGB/vc9Hw8upWtbYCqvbUFB/tb50TGx/4wFKDr69jftQYH+2uPG/eTuJkdWX2Xw8NCFy/Y3dF+4/If8szv+N3LsvGV6/cWyqvfyrUSh4BN59NffNzTw2/lqXq+qPDySd1Zn9c2sJ/ME6Gx0fLKfwTE+E7i/tLJak7ty2T0YgwcP1pUrV9S7d2+tXbvW+sqQIYO+++47rV27VmvWrLnnfrp3764rV67YvTw8stxzu0ctMTFRERFRqlK5onXMZDKpcuWK2rYt3OE227aHq3KVinZjVV9+McX5eHBJiUk6FHVE5Sre7mU1mUwqV7GM9oYb+4HGZPJQxoxehu4zPbv13tmrypWft46ZTCZVfqmitm13/F7Yvi3Cbr4kvVzlBW1PYT4eXErXtir3uLZVucu17eTJMzp/PsZun1myPKby5UuneM6RnJenp555+nFt33fEOmY2m7V931GVKPSkw23+TkhM9uVehn+SipQ+bl3/O0Fnoy+kmLQgOT4TIL1zWQWjW7duevnll/X++++rVq1aGjJkiLy8Uv+hy9vbO9myhe7aHjVy1CRNnTJC4RFR2rlzt9q3ay5f38z6/vt5kqRpU0fp93Pn1avXF5KkMaOnKCxsoTp2bKHly1frrbfqKDi4hFq1/tS6zxw5suuJJ/Iqd+4gSVKhQrd6N6OjY/lWI5VmfztffUd218HIw9q/+6Dead5AmX0ya8ncZZKkfqN6KC46XmOHfCvp1o3h/y436+XlpYDc/ipUrICuX/vLWrFo0/1jbVmzXdG/x8jnMR/VeKOqgp8rpXbvdnHJMaZVo76ZpCmThys8Ikq7du5Ru3bN5OubWdOnz5ckTZkyQufORat37y8lSWPGTtHqVQvUscPHWr48TA3eqq3g4BJq3eb2Sis5cmRXvnx5lOeO905MTBzvnVQaMWqSpjm4tn1nc207d+68ev5zbRs9eorWhC1Up44ttGz5ajX859rW0uba9s3oyerRvb2OHjuhU6fOqn+/rjp3LkaLFq10yTGmVR+8/pJ6j52tYk/nU/ECT2rmsvX6K+GG6r5063k8PcfMUqBfNnV493VJUqXgYpqxdJ2KPJVXzxZ8Umej4zV23nK9GFzMmmh8PX2RKpUtptz+foq7dEXj569QBg+TXq1YxmXHmRbxmQDpmUtv8i5XrpzCw8PVpk0blS1bVrNmzXLb5MAICxYsVoC/n/r26aJcuQIUGblfr7/+vmL/uRk1X748MtssRbl12y590Kit+vf/VIMGfqajx06q/pvNtH//YeucWq9X05QpI6x/nj1rvCRpwMCvNXDg3e9fgb1Vi9coe87satG1qXIG+OnI/mNq/14XXfxndZRceYNkMd/+Di8gyF+zVk21/vmDVu/og1bvKHzLbrV8s4MkKYd/DvX7pof8A3Pqzz+u6djB42r3bhft2LBLuH8LFy5RgL+f+vT5RLmCAhQZeUC1an9g897JK7PNudm2LVyNGrdT/35dNWDApzp27JQaNPhIBw7cfu+8/vormjzp9ntk1sxby90OHDRcgwbdfk/h3v69tvWzubbVtLm2PeHg2vZ+o7YacJdr21fDxsnX10cTxg1V9uxZtXnzTtWs9X6ye+5wdzWeK61LV//UuPkrFH/5qgrnz6txPVpYb/yOjr8kD5t/d5vXf0UmkzR27nLFXryiHFl9VSm4mNq+U9M6J+biFXUbNUOX/7imHFkfU+kiT2vG5x3ll/WxR358aRmfCdxbWlmtyV25zXMw5s6dq44dOyouLk579+5V0aIP/lRdd30OBm5x1+dgwH2fgwH3fg4G3Pc5GHDf52DAvZ+D0Th/fVeHkKLvT/3g6hDuyW2WqX377bdVsWJFhYeH68knHfeGAgAAAHBvbpNgSNLjjz+uxx9/3NVhAAAA4D/M7B4NPmmW2z3JGwAAAEDaRYIBAAAAwDBu1SIFAAAAuBoNUs6hggEAAADAMCQYAAAAAAxDixQAAABgw0yTlFOoYAAAAAAwDAkGAAAAAMPQIgUAAADYsNAi5RQqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2zK4OII2jggEAAADAMCQYAAAAQDo1duxY5c+fX5kyZVJISIh27NiR4txJkybphRdeUI4cOZQjRw5VrVr1rvNTQoIBAAAA2DDL4rav1Jg3b546d+6svn37KiIiQiVLllT16tUVGxvrcP66dev0zjvvaO3atdq6davy5cunatWq6ffff0/V7yXBAAAAANKh4cOHq3nz5mrSpImKFi2qCRMmyMfHR1OnTnU4f9asWWrdurVKlSqlIkWKaPLkyTKbzQoLC0vV7yXBAAAAANKIhIQEXb161e6VkJCQbN6NGzcUHh6uqlWrWsc8PDxUtWpVbd269b5+1/Xr15WYmCg/P79UxUiCAQAAANiwuPH/hgwZomzZstm9hgwZkuwY4uPjdfPmTQUFBdmNBwUFKTo6+r7+Hj777DPlyZPHLkm5HyxTCwAAAKQR3bt3V+fOne3GvL29Df89X3zxhebOnat169YpU6ZMqdqWBAMAAABII7y9ve8rofD391eGDBkUExNjNx4TE6NcuXLdddthw4bpiy++0OrVq1WiRIlUx0iLFAAAAGDD7Mav+5UxY0YFBwfb3aD97w3bFSpUSHG7oUOHauDAgVqxYoXKli2bit94GxUMAAAAIB3q3LmzGjdurLJly6p8+fIaOXKkrl27piZNmkiSGjVqpLx581rv4fjyyy/Vp08fzZ49W/nz57feq/HYY4/pscceu+/fS4IBAAAApEMNGzZUXFyc+vTpo+joaJUqVUorVqyw3vh95swZeXjcbmgaP368bty4oTfffNNuP3379lW/fv3u+/eSYAAAAAA2LJbUPdDOnbVt21Zt27Z1+LN169bZ/fnUqVOG/E7uwQAAAABgGBIMAAAAAIahRQoAAACwYVb6aZFyBSoYAAAAAAxDggEAAADAMLRIAQAAADZS80A7JEcFAwAAAIBhSDAAAAAAGIYWKQAAAMCGhVWknEIFAwAAAIBhSDAAAAAAGIYWKQAAAMAGD9pzDhUMAAAAAIYhwQAAAABgGFqkAAAAABsWCy1SzqCCAQAAAMAwJBgAAAAADEOLFAAAAGDD7OoA0jgqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2LDxozylUMAAAAAAYhgQDAAAAgGFokQIAAABsmGmRcgoVDAAAAACGIcEAAAAAYBhapAAAAAAbFgstUs6gggEAAADAMCQYAAAAAAxDixQAAABgg1WknEMFAwAAAIBhSDAAAAAAGIYWKQAAAMCGhRYpp6TLBIP/JNzbyT+jXR0CUuCdwcvVISAF180Jrg4Bd5E9pKWrQ0AKrv+2ztUhAP85tEgBAAAAMEy6rGAAAAAAD8rMg/acQgUDAAAAgGFIMAAAAAAYhhYpAAAAwAYNUs6hggEAAADAMCQYAAAAAAxDixQAAABgw0yTlFOoYAAAAAAwDAkGAAAAAMPQIgUAAADYoEXKOVQwAAAAABiGBAMAAACAYWiRAgAAAGxYLLRIOYMKBgAAAADDkGAAAAAAMAwtUgAAAIANVpFyDhUMAAAAAIYhwQAAAABgGFqkAAAAABsWWqScQgUDAAAAgGFIMAAAAAAYhhYpAAAAwAYP2nMOFQwAAAAAhiHBAAAAAGAYWqQAAAAAGzxozzlUMAAAAAAYhgQDAAAAgGFokQIAAABssIqUc6hgAAAAADAMCQYAAAAAw9AiBQAAANhgFSnnUMEAAAAAYBgSDAAAAACGoUUKAAAAsGGhRcopVDAAAAAAGIYEAwAAAIBhaJECAAAAbJh50J5TqGAAAAAAMAwJBgAAAADD0CIFAAAA2GAVKedQwQAAAABgGBIMAAAAAIYhwQAAAABgGO7BAAAAAGywTK1zqGAAAAAAMAwJBgAAAADD0CIFAAAA2GCZWudQwQAAAABgGBIMAAAAAIahRQoAAACwwSpSzqGCAQAAAMAwJBgAAAAADEOLFAAAAGCDVaScQwUDAAAAgGFIMAAAAAAYhhYpAAAAwAarSDmHCgYAAAAAw5BguECrlo117Mg2/Xn1uLZsWqJyZUvddX79+q9r3971+vPqce2OWK1Xa1RJNqdf3y46ezpCf1w5ppXL56pAgaceUvTpW9Pm7yli7xr9FrtXK9csUOngEnedX7tuDW3dtUK/xe7Vhq1LVLVaJbufjx7/heKvHrF7zftx8sM8hHSr+ccfaO+BDYq9cFBr1v2o4Hucm7pvvKpdEasUe+Ggtu5YrmrVX7L7efceHbQrYpXOx+7T6d92a9EvM1S2bMmHeATpG9c199WiRSMdPrxZly8f0YYNi+7533m9ejUVGblGly8f0a5dv6p69crWn3l6emrQoO7atetXXbhwSCdO7NSUKSOUO3fQwz6MdGnOj7+oWoOmKvPyG3rn487ae+BwinMTk5I0ftoc1Wj4kcq8/IbqfdhWm7aH283ZtWef2nzWX5XrNlLxF15X2IatD/sQAIdIMB6xBg1qa9hXfTVw0HCVC6mhyKgDWrZ0lgICcjqcXyG0rGbNGKtp0+aobPnqWrx4pX5YOEXFihW2zunapbXatmmq1m276bmKtXTt+nUt+2WWvL29H9VhpQt1672mgYO766svxqjKC3W1f+8hLfhxivz9/RzOL1e+tL6dOlyzpi9Q5Yp1tWzpak2fPVZFniloN2/1qg0qWuA56+vjpp0fxeGkK/Xq19TgL3roiyHf6IXna2nv3oP6cdH38k/hfVM+pIymfjdK06fPV8XnXtfSJb9q9twJeqZoIeucY8dOqssn/VSh/Kuq/spbOnP6N/20eLpypnC+kTKua+7rzTdraejQ3vr885EKDa2pvXsPasmSmSmem9DQYE2fPlrffTdPISGvacmSlVqwYJKK/vPe8fHJrNKli2vIkG8UGvqa3n77YxUs+LQWLpzyKA8rXVgetkFDx0xWqw/f0YLJo1S4wFNq8UkfXbh02eH80ZNmaMHi5erRsYUWzRivt+q8pg49PtfBI8etc/76+28VLvC0enZu+YiOIv2yuPH/0gKTxZL+msw8M+Z1dQgp2rJpiXbuilSHjr0kSSaTSadO7NTYcdM09KuxyebPnjVevj4+qvNGY+vY5o1LtCdyv9q07SZJOns6QiNGTtTwERMlSVmzZtG53/ao6UedNH/+4kdwVKmTPZOvq0NwaOWaBdodsVfdugyQdOvcRB3coEkTZ+ibEd8mmz952kj5+GbWu2+1sI6tCJuvfVEH1aVTX0m3KhjZsmVVo3dbP5qDcNKNm0muDsGhNet+VER4lLp80k/SrXNz8MhmTZwwXSO+npBs/rTvv5Gvr4/eevMj61jY2h8UFXVQnTr0cvg7smR5TL9HR6lWzfe1ft2Wh3IczriemODqEFLEdU3y9Mjg6hAc2rBhkcLDI9WpUx9Jt87NsWPbNX78dxo2bFyy+TNmjJWvr4/q1WtiHVu//mdFRR1Qu3Y9HP6O4OAS2rz5FxUsGKqzZ889nANxwh9nwlwdgkPvfNxZxZ8pqJ6dWkmSzGazqtb/UO/Wr6WP3m+QbH7luo30caO39E69161jHXsNlnfGjPqyT5dk84u/8LpGfd5TL79Y4eEdhJO8Agvee5KLPO1f2tUhpOhE/G5Xh3BPVDAeIS8vL5UpU0JhazZaxywWi8LWbFJoaLDDbUJDgu3mS9Kvq9ZZ5z/11BPKnTtIYWs2WX9+9eof2rFjt0JDHO8TyXl5ealkqWJav/b2B0uLxaL167aoXPlSDrcpW75Usg+ia8M2qWx5+4vS8xXL6+DxrdoWvkJfDe+nHH7ZjQ4/XfPy8lKp0sW1du1m65jFYtG6tZtVvrzjfwDKh5TROpv5khS2eqPKhzie7+XlpQ+bvq3Ll69q796DxgX/H8B1zX3dOjfPao3N36PFYtHatZsUElLG4TahoWXs5kvS6tUbUpwvSdmyZZXZbNbly1eNCfw/IDExUQeOHFNocCnrmIeHh0LLllLk/kMOt7mRmKiMGTPajXlnzKjdew88zFCBB+JWq0hdu3ZN8+fP17Fjx5Q7d2698847ypnTcRk3LfL395Onp6diY+LtxmNj41Sk8P8cbpMrV4BiYuPsxmJi4pUrKODWz4MC/xm7Y05svHLlCjQq9HQvZ84c8vT0VFyc/bmJi41XwUJPO9wmMMhfcbF3nst4BQb5W/+8ZvVGLV38q06f/k35n3pCvfp21rwfJqvGy2/JbDYbfyDpkPXcOPi7LlTI8fsmKMhfsQ7mB/3zvvlXjRpVNPX7UfLxyazo6FjVrdVIFy9cMvYA0jmua+7Lem7ueC/ExNztvROg2GTnJi7Ze+df3t7eGjSou+bPX6Q//vjTmMD/Ay5duaqbN83KeccXTjlzZNfJ07853Ob58mU0fd7PKluymPLlza1t4ZEK27BVN803H0HE/z0WC/9GO8OlCUbRokW1adMm+fn56ezZs3rxxRd16dIlFSpUSMePH9fAgQO1bds2PfVUyjf2JSQkKCHBvnXAYrHIZDI97PCBe/rph6XW/3/wwBEd2H9Y4VFhev6FEG1cz813rrZhw1ZVrPC6cubMocZN3tZ3M0arykv1FB93wdWhAW7P09NTs2aNk8kktWvX09XhpHvd2n+sfkNHq9b7rWQySfny5Fbd16rqp6WrXB0akIxLW6QOHTqkpKRbPd/du3dXnjx5dPr0ae3YsUOnT59WiRIl1LPn3S9aQ4YMUbZs2exeFvMfjyL8VIuPv6ikpCS7b7glKTAwQNF3fFP3r+joOAUF2n9zFBTkb50fHRP7z9gdcwL9FR0da1To6d6FC5eUlJSkgAD7cxMQ6K/YFM5NbEy8AgLvPJf+yb7JtXX61FnFx1/U008/4XzQ/xHWc+Pg7/rOb7j/FRMTr8D7mH/9+l86ceK0du7co7atu+lm0k01avyWsQeQznFdc1/Wc3PHeyEo6G7vnTgFJjs3Acnm/5tcPPFEXtWs+R7Vi1TKkS2rMmTw0IWLl+3GL1y6LP+cORxu45cjm74Z0ks7f12oXxdM1ZJZE+STOZMez5PrEUQMpI7b3IOxdetW9evXT9myZZMkPfbYY+rfv782bdp01+26d++uK1eu2L1MHlkeRciplpiYqIiIKFWpXNE6ZjKZVKVyRW3bFu5wm23bw1WlSkW7saovv2idf/LkGZ0/H2O3zyxZHlP58qW1bbvjfSK5xMRERe7Zrxdfun0znMlk0ouVKmjnjj0Ot9m1Y49erGR/81ylys9p146Ub77KnSdIfn7ZFRPt+B93JJeYmKg9u/fppZees46ZTCZVeuk57Ujh73rH9ghVspkvSZWrPK8d2+9+Y5yHh0ned/Q44+64rrmvW+dmrypXft46ZjKZ9NJLz2v79giH22zbFmE3X5KqVKloN//f5KJAgaf02mvv6uIdH5Jxb15eXipaqIC2h0dax8xms7aHR6pksSJ33dbbO6OCAvyVdPOmVq3fosoVQx52uP9JZlnc9pUWuPwejH9bmf7++2/lzp3b7md58+ZVXNzdP4h5e3snW7bQndujRoyapGlTRig8Iko7d+5W+3bN5eubWd99P0+SNG3qKJ07d149e30hSRo9eorWhC1Up44ttGz5ajV8q46Cg0uoZetPrfv8ZvRk9ejeXkePndCpU2fVv19XnTsXo0WLVrrkGNOq8WOmacyEL7Vn9z5F7IpSy9aN5eOTWXNm/iBJGjtxqM6fi9Gg/l9LkiaO/16Ll89U67ZN9evKdar3Zk2VKl1cndv3liT5+vqoa7e2WrJ4pWJj4pX/qSfUb0BXnTxxWmvCNqYYB5IbM3qKJnw7TLt379WuXZFq3aaJfHx8NHPGQknSxEnDdO5cjPr3/UqSNH7cd1q+co7atm+mlSvW6s03a6l0mWfV/p82Dh+fzOryaRstX7pa0dGxypnTT81bfKDceXLpp5+Wuew40yqua+7rm28ma/LkrxURsVc7d+5Ru3bN5Ovro+nT50uSpkwZoXPnotW795eSpLFjp2rVqvnq0KG5li9fo7feqq3g4BJq0+bW6l6enp6aM2eCSpcurjfeaKIMGTJYK00XL15WYmKiaw40DWrUsK56Dh6hYkUKqvgzhTRzwSL99dffqvtaVUlS90FfK9A/pzq1/FCSFLX/sGLiL6hIwacVGxevcVNny2I2q+m79a37vH79L535/bz1z7+fj9GhoyeULetjyh3E/Ut4dFyeYLz88svy9PTU1atXdfjwYRUvXtz6s9OnT6erm7wlacGCxQrw91O/Pl2UK1eAIiP3q+br71tvwnsiXx67m3+3btul9xu11YD+n2rQwM909NhJ1X+zmfbvv/0wnq+GjZOvr48mjBuq7NmzavPmnapZ6/1k96bg7n7+cZly+vupW4/2CgwK0L69B/VW/WaK+6cf//HHc9udm507dqtFs0/Uo3dH9ezbWSeOn1Kjd9vo0MGjkqSbN2+qaPHCavjuG8qWLYuiz8dq3ZrNGjJopG7c4B/h1Pjxh6Xy9/dTj16dFBTkr71RB1W/7ofWG78ff9z+fbNje4SaNemo3n0+Ud9+XXT8+Cm9+3ZLHTxwRNKtc1Oo0P/07nv1lDNnDl28eFkR4VGq8UpD6/nD/eO65r4WLlwif38/9enTWUFBAYqMPKDatT+wnpt8d5ybbdvC1bhxe/Xr10UDBnyqY8dOqUGD5jrwz3snb95cqlWrmiRp5077ZK9atbe0YcO2R3Rkad+rL7+oS5evaMyUmYq/eElFCjytCcMGyN/vVovU+Zg4eZhuN5ok3Lih0ZNm6Lfz0fLJnFkvhAZrSO9PlDXLY9Y5+w4fVdP2t5cTHjrm1oNd69R4WZ/37PSIjgxw8XMw+vfvb/fn0NBQVa9e3frnrl276rffftOcOXNStV93fg4G3Pc5GHDf52DAvZ+DAfd9Dgbc9zkYcO/nYDzh96yrQ0jRmYt7XR3CPfGgPTxyJBjuiwTDfZFguDcSDPdFguG+SDAeTFpIMNzmJm8AAAAAaZ/L78EAAAAA3ElaWa3JXVHBAAAAAGAYEgwAAAAAhqFFCgAAALCRDtdAeqSoYAAAAAAwDAkGAAAAAMPQIgUAAADYMNMi5RQqGAAAAAAMQ4IBAAAAwDC0SAEAAAA2LDxozylUMAAAAAAYhgQDAAAAgGFokQIAAABs8KA951DBAAAAAGAYEgwAAAAAhqFFCgAAALBhZhUpp1DBAAAAAGAYEgwAAAAAhqFFCgAAALDBKlLOoYIBAAAAwDAkGAAAAAAMQ4sUAAAAYMNMi5RTqGAAAAAAMAwJBgAAAADD0CIFAAAA2GAVKedQwQAAAABgGBIMAAAAAIahRQoAAACwYRYtUs6gggEAAADAMCQYAAAAAAxDixQAAABgg1WknEMFAwAAAIBhSDAAAAAAGIYWKQAAAMCGmRYpp1DBAAAAAGAYEgwAAAAAhqFFCgAAALBh4UF7TqGCAQAAAMAwJBgAAAAADEOLFAAAAGCDVaScQwUDAAAAgGFIMAAAAAAYhhYpAAAAwIaFFimnUMEAAAAAYBgSDAAAAACGoUUKAAAAsMGD9pxDBQMAAACAYUgwAAAAABiGFikAAADABqtIOYcKBgAAAADDkGAAAAAAMAwtUgAAAIANWqScQwUDAAAAgGFIMAAAAIB0auzYscqfP78yZcqkkJAQ7dix467zFyxYoCJFiihTpkx69tlntWzZslT/ThIMAAAAwIbFjV+pMW/ePHXu3Fl9+/ZVRESESpYsqerVqys2Ntbh/C1btuidd95Rs2bNtHv3btWtW1d169bVvn37UvV7TZZ02GTmmTGvq0PAXWTP5OvqEJCCGzeTXB0CUnA9McHVIeAuPD0yuDoEpOCPM2GuDgEp8Aos6OoQUuTOnyWTbvx+33NDQkJUrlw5jRkzRpJkNpuVL18+tWvXTt26dUs2v2HDhrp27Zp++eUX61hoaKhKlSqlCRMm3PfvpYIBAAAApBEJCQm6evWq3SshIfmXUDdu3FB4eLiqVq1qHfPw8FDVqlW1detWh/veunWr3XxJql69eorzU5IuV5FKTWbn7hISEjRkyBB1795d3t7erg4HNjg37o3z4744N+6Lc+PeOD+Pjjt/luzXr5/69+9vN9a3b1/169fPbiw+Pl43b95UUFCQ3XhQUJAOHTrkcN/R0dEO50dHR6cqRioYbi4hIUH9+/d3mJnCtTg37o3z4744N+6Lc+PeOD+QpO7du+vKlSt2r+7du7s6LDvpsoIBAAAApEfe3t73VcHy9/dXhgwZFBMTYzceExOjXLlyOdwmV65cqZqfEioYAAAAQDqTMWNGBQcHKyzs9kIHZrNZYWFhqlChgsNtKlSoYDdfklatWpXi/JRQwQAAAADSoc6dO6tx48YqW7asypcvr5EjR+ratWtq0qSJJKlRo0bKmzevhgwZIknq0KGDKlWqpK+//lo1a9bU3LlztWvXLn377bep+r0kGG7O29tbffv25WYuN8S5cW+cH/fFuXFfnBv3xvlBajVs2FBxcXHq06ePoqOjVapUKa1YscJ6I/eZM2fk4XG7oem5557T7Nmz1atXL/Xo0UMFCxbUzz//rOLFi6fq96bL52AAAAAAcA3uwQAAAABgGBIMAAAAAIYhwQAAAABgGBIMAAAAAIYhwXBjY8eOVf78+ZUpUyaFhIRox44drg4JkjZs2KBatWopT548MplM+vnnn10dEv4xZMgQlStXTlmyZFFgYKDq1q2rw4cPuzos/GP8+PEqUaKEsmbNqqxZs6pChQpavny5q8OCA1988YVMJpM6duzo6lD+8/r16yeTyWT3KlKkiKvDAu6KBMNNzZs3T507d1bfvn0VERGhkiVLqnr16oqNjXV1aP95165dU8mSJTV27FhXh4I7rF+/Xm3atNG2bdu0atUqJSYmqlq1arp27ZqrQ4Okxx9/XF988YXCw8O1a9cuValSRXXq1NH+/ftdHRps7Ny5UxMnTlSJEiVcHQr+UaxYMZ0/f9762rRpk6tDAu6KZWrdVEhIiMqVK6cxY8ZIuvXkxXz58qldu3bq1q2bi6PDv0wmk3766SfVrVvX1aHAgbi4OAUGBmr9+vV68cUXXR0OHPDz89NXX32lZs2auToUSPrzzz9VpkwZjRs3ToMGDVKpUqU0cuRIV4f1n9avXz/9/PPP2rNnj6tDAe4bFQw3dOPGDYWHh6tq1arWMQ8PD1WtWlVbt251YWRA2nLlyhVJtz7Ewr3cvHlTc+fO1bVr11ShQgVXh4N/tGnTRjVr1rT79weud/ToUeXJk0dPP/30/9u725AqzweO49+DddyZnjqz7DmdTXIGPalMpFZJD7MXYsloNLeOPUGl68G1LV+sh0XZXgi1AhUqLcpKKk00Js6h1qDRAycsyqUYFbOyoug01PK4F6Pz76xs2f/YbfX7gOC5z33u6+f9Qvid67q4SUpK4sqVK0ZHEnkuPcm7G7p16xZtbW3upyw+1r9/fy5evGhQKpHXi8vlYvny5YwbN67TTyCVrlNTU0NMTAzNzc34+/tTWFjIiBEjjI4lwP79+zlz5gwnT540Ooo8ITo6mry8PMLCwmhsbGTdunV8/PHHnDt3DqvVanQ8kWdSwRCRN1JKSgrnzp3TWuVuJiwsDIfDwb179zh48CB2u52qqiqVDINdvXqVZcuWUV5ezjvvvGN0HHnC9OnT3b+PGjWK6OhogoODKSgo0NJC6bZUMLqhvn374uPjw40bNzyO37hxgwEDBhiUSuT1kZqaSklJCdXV1QwZMsToOPIEs9lMaGgoAJGRkZw8eZItW7aQk5NjcLK32+nTp7l58yYRERHuY21tbVRXV7Nt2zZaWlrw8fExMKE8ZrPZGD58OHV1dUZHEemQ9mB0Q2azmcjISCoqKtzHXC4XFRUVWqss8hzt7e2kpqZSWFjIr7/+SkhIiNGR5D+4XC5aWlqMjvHWmzx5MjU1NTgcDvdPVFQUSUlJOBwOlYtuxOl0Ul9fz8CBA42OItIhzWB0U2lpadjtdqKiovjoo4/YvHkzDx48YO7cuUZHe+s5nU6Pb44aGhpwOBwEBAQQFBRkYDJJSUkhPz+fI0eOYLVauX79OgC9e/fGYrEYnE7S09OZPn06QUFB3L9/n/z8fCorKykrKzM62lvParU+tVfJz8+PPn36aA+TwVauXEl8fDzBwcH8+eefrFmzBh8fH2bPnm10NJEOqWB0U5999hlNTU2sXr2a69evM2bMGH7++eenNn7Lq3fq1CliY2Pdr9PS0gCw2+3k5eUZlErgnwe5AUyaNMnjeG5uLsnJya8+kHi4efMmc+bMobGxkd69ezNq1CjKysqYOnWq0dFEuq1r164xe/Zsbt++TWBgIOPHj+fEiRMEBgYaHU2kQ3oOhoiIiIiIeI32YIiIiIiIiNeoYIiIiIiIiNeoYIiIiIiIiNeoYIiIiIiIiNeoYIiIiIiIiNeoYIiIiIiIiNeoYIiIiIiIiNeoYIiIdDPJycnMmDHD/XrSpEksX778leeorKzEZDJx9+7dVz62iIi8vlQwREReUHJyMiaTCZPJhNlsJjQ0lB9++IFHjx516biHDx9m/fr1L3SuSoGIiBith9EBREReJ3FxceTm5tLS0sLRo0dJSUmhZ8+epKene5zX2tqK2Wz2ypgBAQFeuY6IiMiroBkMEZFO8PX1ZcCAAQQHB7N48WKmTJlCcXGxe1nThg0bGDRoEGFhYQBcvXqVWbNmYbPZCAgIICEhgcuXL7uv19bWRlpaGjabjT59+vDtt9/S3t7uMea/l0i1tLTw3XffMXToUHx9fQkNDWXHjh1cvnyZ2NhYAN577z1MJhPJyckAuFwuMjIyCAkJwWKxMHr0aA4ePOgxztGjRxk+fDgWi4XY2FiPnCIiIi9KBUNE5P9gsVhobW0FoKKigtraWsrLyykpKeHhw4d88sknWK1Wjh07xm+//Ya/vz9xcXHuz2RmZpKXl8fOnTs5fvw4d+7cobCw8Lljzpkzh3379vHTTz9x4cIFcnJy8Pf3Z+jQoRw6dAiA2tpaGhsb2bJlCwAZGRns3r2b7Oxszp8/z4oVK/jiiy+oqqoC/ilCiYmJxMfH43A4WLBgAatWreqq2yYiIm8wLZESEXkJ7e3tVFRUUFZWxldffUVTUxN+fn5s377dvTRqz549uFwutm/fjslkAiA3NxebzUZlZSXTpk1j8+bNpKenk5iYCEB2djZlZWUdjvvHH39QUFBAeXk5U6ZMAWDYsGHu9x8vp+rXrx82mw34Z8Zj48aN/PLLL8TExLg/c/z4cXJycpg4cSJZWVl88MEHZGZmAhAWFkZNTQ0//vijF++aiIi8DVQwREQ6oaSkBH9/fx4+fIjL5eLzzz9n7dq1pKSkMHLkSI99F2fPnqWurg6r1epxjebmZurr67l37x6NjY1ER0e73+vRowdRUVFPLZN6zOFw4OPjw8SJE184c11dHX/99RdTp071ON7a2srYsWMBuHDhgkcOwF1GREREOkMFQ0SkE2JjY8nKysJsNjNo0CB69Pjfv1E/Pz+Pc51OJ5GRkezdu/ep6wQGBr7U+BaLpdOfcTqdAJSWljJ48GCP93x9fV8qh4iISEdUMEREOsHPz4/Q0NAXOjciIoIDBw7Qr18/evXq9cxzBg4cyO+//86ECRMAePToEadPnyYiIuKZ548cORKXy0VVVZV7idSTHs+gtLW1uY+NGDECX19frly50uHMR3h4OMXFxR7HTpw48d9/pIiIyL9ok7eISBdJSkqib9++JCQkcOzYMRoaGqisrGTp0qVcu3YNgGXLlrFp0yaKioq4ePEiS5Ysee4zLN5//33sdjvz5s2jqKjIfc2CggIAgoODMZlMlJSU0NTUhNPpxGq1snLlSlasWMGuXbuor6/nzJkzbN26lV27dgGwaNEiLl26xDfffENtbS35+fnk5eV19S0SEZE3kAqGiEgXeffdd6muriYoKIjExETCw8OZP38+zc3N7hmNr7/+mi+//BK73U5MTAxWq5WZM2c+97pZWVl8+umnLFmyhA8//JCFCxfy4MEDAAYPHsy6detYtWoV/fv3JzU1FYD169fz/fffk5GRQXh4OHFxcZSWlhISEgJAUFAQhw4doqioiNGjR5Odnc3GjRu78O6IiMibytTe0U5CERERERGRTtIMhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeI0KhoiIiIiIeM3fnjWy3ozXsKgAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "# Normalise\n",
    "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show(block=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34242624",
   "metadata": {},
   "source": [
    "### Combined MCI groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "bad8a5cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('DTI_with_group_comb.csv')\n",
    "# remove unnecessary columns + confounding\n",
    "df = df.drop(['IMAGEUID', 'COLPROT', 'RID', 'VISCODE', \n",
    "                          'VISCODE2', 'EXAMDATE', 'VERSION', 'MANUFACTURER',\n",
    "                         'RUNDATE', 'STATUS', 'QC', 'update_stamp', 'Age', 'VOLUMES', 'Sex'], axis=1)\n",
    "# remove rows with empty values\n",
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "a52f9c65",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.iloc[:, df.columns != 'Research Group'].values\n",
    "y = df.iloc[:, -1].values\n",
    "X_resampled, y_resampled = SVMSMOTE().fit_resample(X, y)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3, random_state=23)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "0099e40b",
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "0595b239",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate:  0.05\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.799\n",
      "Learning rate:  0.075\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.810\n",
      "Learning rate:  0.1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.792\n",
      "Learning rate:  0.25\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.808\n",
      "Learning rate:  0.5\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.782\n",
      "Learning rate:  0.75\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.775\n",
      "Learning rate:  1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.794\n"
     ]
    }
   ],
   "source": [
    "lr_list = [0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1]\n",
    "\n",
    "for learning_rate in lr_list:\n",
    "    gb_clf = GradientBoostingClassifier(n_estimators=2000, learning_rate=learning_rate, max_depth=5, random_state=23)\n",
    "    gb_clf.fit(X_train, y_train)\n",
    "\n",
    "    print(\"Learning rate: \", learning_rate)\n",
    "    print(\"Accuracy score (training): {0:.3f}\".format(gb_clf.score(X_train, y_train)))\n",
    "    print(\"Accuracy score (validation): {0:.3f}\".format(gb_clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "cfd0e3d5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[ 97   1   3]\n",
      " [  3 115  30]\n",
      " [  2  42 134]]\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          AD       0.95      0.96      0.96       101\n",
      "          CN       0.73      0.78      0.75       148\n",
      "         MCI       0.80      0.75      0.78       178\n",
      "\n",
      "    accuracy                           0.81       427\n",
      "   macro avg       0.83      0.83      0.83       427\n",
      "weighted avg       0.81      0.81      0.81       427\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gb_clf2 = GradientBoostingClassifier(n_estimators=2000, learning_rate=0.075, max_depth=5, random_state=23)\n",
    "gb_clf2.fit(X_train, y_train)\n",
    "predictions = gb_clf2.predict(X_test)\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, predictions))\n",
    "\n",
    "print(\"Classification Report\")\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "37e3bce6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, predictions)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(8, 8))\n",
    "ax.imshow(cm)\n",
    "ax.grid(False)\n",
    "ax.set_xlabel('Predicted outputs', fontsize=12, color='black')\n",
    "ax.set_ylabel('Actual outputs', fontsize=12, color='black')\n",
    "ax.xaxis.set(ticks=range(3))\n",
    "ax.yaxis.set(ticks=range(3))\n",
    "for i in range(3):\n",
    "    for j in range(3):\n",
    "        ax.text(j, i, cm[i, j], ha='center', va='center', color='white')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "24cfe1e3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "# Normalise\n",
    "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show(block=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8fff28a",
   "metadata": {},
   "source": [
    "# SMOTE + ENN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "095805bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('DTI_with_group.csv')\n",
    "# remove unnecessary columns + confounding\n",
    "df = df.drop(['IMAGEUID', 'COLPROT', 'RID', 'VISCODE', \n",
    "                          'VISCODE2', 'EXAMDATE', 'VERSION', 'MANUFACTURER',\n",
    "                         'RUNDATE', 'STATUS', 'QC', 'update_stamp', 'Age', 'VOLUMES', 'Sex'], axis=1)\n",
    "# remove rows with empty values\n",
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "b8941dc9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.combine import SMOTEENN\n",
    "smote_enn = SMOTEENN(random_state=4)\n",
    "X = df.iloc[:, df.columns != 'Research Group'].values\n",
    "y = df.iloc[:, -1].values\n",
    "X_resampled, y_resampled = smote_enn.fit_resample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "e39ea046",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3, random_state=23)\n",
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "7018e5f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate:  0.05\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.965\n",
      "Learning rate:  0.075\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.969\n",
      "Learning rate:  0.1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.968\n",
      "Learning rate:  0.25\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.964\n",
      "Learning rate:  0.5\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.961\n",
      "Learning rate:  0.75\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.955\n",
      "Learning rate:  1\n",
      "Accuracy score (training): 0.999\n",
      "Accuracy score (validation): 0.937\n"
     ]
    }
   ],
   "source": [
    "lr_list = [0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1]\n",
    "\n",
    "for learning_rate in lr_list:\n",
    "    gb_clf = GradientBoostingClassifier(n_estimators=2000, learning_rate=learning_rate, max_depth=5, random_state=23)\n",
    "    gb_clf.fit(X_train, y_train)\n",
    "\n",
    "    print(\"Learning rate: \", learning_rate)\n",
    "    print(\"Accuracy score (training): {0:.3f}\".format(gb_clf.score(X_train, y_train)))\n",
    "    print(\"Accuracy score (validation): {0:.3f}\".format(gb_clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "6ad4f9bf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[145   0   0   1   1   1]\n",
      " [  0   2   1   1   3   0]\n",
      " [  1   0 129   2   1   0]\n",
      " [  2   0   0 144   0   0]\n",
      " [  1   0   2   0 118   5]\n",
      " [  0   0   0   0   0 158]]\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          AD       0.97      0.98      0.98       148\n",
      "          CN       1.00      0.29      0.44         7\n",
      "        EMCI       0.98      0.97      0.97       133\n",
      "        LMCI       0.97      0.99      0.98       146\n",
      "         MCI       0.96      0.94      0.95       126\n",
      "         SMC       0.96      1.00      0.98       158\n",
      "\n",
      "    accuracy                           0.97       718\n",
      "   macro avg       0.97      0.86      0.88       718\n",
      "weighted avg       0.97      0.97      0.97       718\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gb_clf2 = GradientBoostingClassifier(n_estimators=2000, learning_rate=0.075, max_depth=5, random_state=23)\n",
    "gb_clf2.fit(X_train, y_train)\n",
    "predictions = gb_clf2.predict(X_test)\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, predictions))\n",
    "\n",
    "print(\"Classification Report\")\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "99d8a6fd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, predictions)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(8, 8))\n",
    "ax.imshow(cm)\n",
    "ax.grid(False)\n",
    "ax.set_xlabel('Predicted outputs', fontsize=12, color='black')\n",
    "ax.set_ylabel('Actual outputs', fontsize=12, color='black')\n",
    "ax.xaxis.set(ticks=range(6))\n",
    "ax.yaxis.set(ticks=range(6))\n",
    "for i in range(6):\n",
    "    for j in range(6):\n",
    "        ax.text(j, i, cm[i, j], ha='center', va='center', color='white')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "44c38ca0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "# Normalise\n",
    "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show(block=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ff70c7de",
   "metadata": {},
   "source": [
    "### Combined MCI groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "3be6ed0f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('DTI_with_group_comb.csv')\n",
    "# remove unnecessary columns + confounding\n",
    "df = df.drop(['IMAGEUID', 'COLPROT', 'RID', 'VISCODE', \n",
    "                          'VISCODE2', 'EXAMDATE', 'VERSION', 'MANUFACTURER',\n",
    "                         'RUNDATE', 'STATUS', 'QC', 'update_stamp', 'Age', 'VOLUMES', 'Sex'], axis=1)\n",
    "# remove rows with empty values\n",
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "02f330a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "smote_enn = SMOTEENN(random_state=4)\n",
    "X = df.iloc[:, df.columns != 'Research Group'].values\n",
    "y = df.iloc[:, -1].values\n",
    "X_resampled, y_resampled = smote_enn.fit_resample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "66a0ac7b",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3, random_state=23)\n",
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "1bbfc8d6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Learning rate:  0.05\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.934\n",
      "Learning rate:  0.075\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.934\n",
      "Learning rate:  0.1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.930\n",
      "Learning rate:  0.25\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.940\n",
      "Learning rate:  0.5\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.934\n",
      "Learning rate:  0.75\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.934\n",
      "Learning rate:  1\n",
      "Accuracy score (training): 1.000\n",
      "Accuracy score (validation): 0.924\n"
     ]
    }
   ],
   "source": [
    "lr_list = [0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1]\n",
    "\n",
    "for learning_rate in lr_list:\n",
    "    gb_clf = GradientBoostingClassifier(n_estimators=2000, learning_rate=learning_rate, max_depth=5, random_state=23)\n",
    "    gb_clf.fit(X_train, y_train)\n",
    "\n",
    "    print(\"Learning rate: \", learning_rate)\n",
    "    print(\"Accuracy score (training): {0:.3f}\".format(gb_clf.score(X_train, y_train)))\n",
    "    print(\"Accuracy score (validation): {0:.3f}\".format(gb_clf.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "474d9978",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[176   0   0]\n",
      " [  0  17  11]\n",
      " [  3   4  91]]\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          AD       0.98      1.00      0.99       176\n",
      "          CN       0.81      0.61      0.69        28\n",
      "         MCI       0.89      0.93      0.91        98\n",
      "\n",
      "    accuracy                           0.94       302\n",
      "   macro avg       0.89      0.85      0.87       302\n",
      "weighted avg       0.94      0.94      0.94       302\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gb_clf2 = GradientBoostingClassifier(n_estimators=2000, learning_rate=0.25, max_depth=5, random_state=23)\n",
    "gb_clf2.fit(X_train, y_train)\n",
    "predictions = gb_clf2.predict(X_test)\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, predictions))\n",
    "\n",
    "print(\"Classification Report\")\n",
    "print(classification_report(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "b512fdc5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, predictions)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(8, 8))\n",
    "ax.imshow(cm)\n",
    "ax.grid(False)\n",
    "ax.set_xlabel('Predicted outputs', fontsize=12, color='black')\n",
    "ax.set_ylabel('Actual outputs', fontsize=12, color='black')\n",
    "ax.xaxis.set(ticks=range(3))\n",
    "ax.yaxis.set(ticks=range(3))\n",
    "for i in range(3):\n",
    "    for j in range(3):\n",
    "        ax.text(j, i, cm[i, j], ha='center', va='center', color='white')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "0cf54ed7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "# Normalise\n",
    "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show(block=False)"
   ]
  }
 ],
 "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": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}