gru_glaukoma.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.layers import Dense, Dropout, GRU, LSTM\n",
    "from keras.models import Sequential\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score, ConfusionMatrixDisplay\n",
    "import plotly.express as px "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('datasets/GlaucomaM.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "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>ag</th>\n",
       "      <th>at</th>\n",
       "      <th>as</th>\n",
       "      <th>an</th>\n",
       "      <th>ai</th>\n",
       "      <th>eag</th>\n",
       "      <th>eat</th>\n",
       "      <th>eas</th>\n",
       "      <th>ean</th>\n",
       "      <th>eai</th>\n",
       "      <th>...</th>\n",
       "      <th>tmt</th>\n",
       "      <th>tms</th>\n",
       "      <th>tmn</th>\n",
       "      <th>tmi</th>\n",
       "      <th>mr</th>\n",
       "      <th>rnf</th>\n",
       "      <th>mdic</th>\n",
       "      <th>emd</th>\n",
       "      <th>mv</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.220</td>\n",
       "      <td>0.354</td>\n",
       "      <td>0.580</td>\n",
       "      <td>0.686</td>\n",
       "      <td>0.601</td>\n",
       "      <td>1.267</td>\n",
       "      <td>0.336</td>\n",
       "      <td>0.346</td>\n",
       "      <td>0.255</td>\n",
       "      <td>0.331</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018</td>\n",
       "      <td>-0.230</td>\n",
       "      <td>-0.510</td>\n",
       "      <td>-0.158</td>\n",
       "      <td>0.841</td>\n",
       "      <td>0.410</td>\n",
       "      <td>0.137</td>\n",
       "      <td>0.239</td>\n",
       "      <td>0.035</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.681</td>\n",
       "      <td>0.475</td>\n",
       "      <td>0.672</td>\n",
       "      <td>0.868</td>\n",
       "      <td>0.667</td>\n",
       "      <td>2.053</td>\n",
       "      <td>0.440</td>\n",
       "      <td>0.520</td>\n",
       "      <td>0.639</td>\n",
       "      <td>0.454</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.014</td>\n",
       "      <td>-0.165</td>\n",
       "      <td>-0.317</td>\n",
       "      <td>-0.192</td>\n",
       "      <td>0.924</td>\n",
       "      <td>0.256</td>\n",
       "      <td>0.252</td>\n",
       "      <td>0.329</td>\n",
       "      <td>0.022</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.979</td>\n",
       "      <td>0.343</td>\n",
       "      <td>0.508</td>\n",
       "      <td>0.624</td>\n",
       "      <td>0.504</td>\n",
       "      <td>1.200</td>\n",
       "      <td>0.299</td>\n",
       "      <td>0.396</td>\n",
       "      <td>0.259</td>\n",
       "      <td>0.246</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.097</td>\n",
       "      <td>-0.235</td>\n",
       "      <td>-0.337</td>\n",
       "      <td>-0.020</td>\n",
       "      <td>0.795</td>\n",
       "      <td>0.378</td>\n",
       "      <td>0.152</td>\n",
       "      <td>0.250</td>\n",
       "      <td>0.029</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.747</td>\n",
       "      <td>0.269</td>\n",
       "      <td>0.476</td>\n",
       "      <td>0.525</td>\n",
       "      <td>0.476</td>\n",
       "      <td>0.612</td>\n",
       "      <td>0.147</td>\n",
       "      <td>0.017</td>\n",
       "      <td>0.044</td>\n",
       "      <td>0.405</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.035</td>\n",
       "      <td>-0.449</td>\n",
       "      <td>-0.217</td>\n",
       "      <td>-0.091</td>\n",
       "      <td>0.746</td>\n",
       "      <td>0.200</td>\n",
       "      <td>0.027</td>\n",
       "      <td>0.078</td>\n",
       "      <td>0.023</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.990</td>\n",
       "      <td>0.599</td>\n",
       "      <td>0.686</td>\n",
       "      <td>1.039</td>\n",
       "      <td>0.667</td>\n",
       "      <td>2.513</td>\n",
       "      <td>0.543</td>\n",
       "      <td>0.607</td>\n",
       "      <td>0.871</td>\n",
       "      <td>0.492</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.105</td>\n",
       "      <td>0.084</td>\n",
       "      <td>-0.012</td>\n",
       "      <td>-0.054</td>\n",
       "      <td>0.977</td>\n",
       "      <td>0.193</td>\n",
       "      <td>0.297</td>\n",
       "      <td>0.354</td>\n",
       "      <td>0.034</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 63 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      ag     at     as     an     ai    eag    eat    eas    ean    eai  ...  \\\n",
       "0  2.220  0.354  0.580  0.686  0.601  1.267  0.336  0.346  0.255  0.331  ...   \n",
       "1  2.681  0.475  0.672  0.868  0.667  2.053  0.440  0.520  0.639  0.454  ...   \n",
       "2  1.979  0.343  0.508  0.624  0.504  1.200  0.299  0.396  0.259  0.246  ...   \n",
       "3  1.747  0.269  0.476  0.525  0.476  0.612  0.147  0.017  0.044  0.405  ...   \n",
       "4  2.990  0.599  0.686  1.039  0.667  2.513  0.543  0.607  0.871  0.492  ...   \n",
       "\n",
       "     tmt    tms    tmn    tmi     mr    rnf   mdic    emd     mv   Class  \n",
       "0 -0.018 -0.230 -0.510 -0.158  0.841  0.410  0.137  0.239  0.035  normal  \n",
       "1 -0.014 -0.165 -0.317 -0.192  0.924  0.256  0.252  0.329  0.022  normal  \n",
       "2 -0.097 -0.235 -0.337 -0.020  0.795  0.378  0.152  0.250  0.029  normal  \n",
       "3 -0.035 -0.449 -0.217 -0.091  0.746  0.200  0.027  0.078  0.023  normal  \n",
       "4 -0.105  0.084 -0.012 -0.054  0.977  0.193  0.297  0.354  0.034  normal  \n",
       "\n",
       "[5 rows x 63 columns]"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Class\n",
      "normal      98\n",
      "glaucoma    98\n",
      "Name: count, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(data[\"Class\"].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = data['Class'].apply(lambda x: 0 if x == 'normal' else 1).to_numpy()\n",
    "x = data.drop(['Class'], axis=1).to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((196, 62), (196,))"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## GRU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_gru = Sequential()\n",
    "model_gru.add(GRU(62, return_sequences=True, input_shape=(62, 1)))\n",
    "model_gru.add(Dropout(0.2))\n",
    "model_gru.add(GRU(50, return_sequences=True))\n",
    "model_gru.add(Dropout(0.2))\n",
    "model_gru.add(GRU(50, return_sequences=False))\n",
    "model_gru.add(Dropout(0.2))\n",
    "model_gru.add(Dense(1, activation = \"sigmoid\"))\n",
    "\n",
    "model_gru.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/12\n",
      "6/6 [==============================] - 3s 120ms/step - loss: 0.6927 - accuracy: 0.5064 - val_loss: 0.6881 - val_accuracy: 0.4750\n",
      "Epoch 2/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6885 - accuracy: 0.5128 - val_loss: 0.6832 - val_accuracy: 0.4750\n",
      "Epoch 3/12\n",
      "6/6 [==============================] - 0s 39ms/step - loss: 0.6846 - accuracy: 0.5064 - val_loss: 0.6746 - val_accuracy: 0.4750\n",
      "Epoch 4/12\n",
      "6/6 [==============================] - 0s 37ms/step - loss: 0.6793 - accuracy: 0.5449 - val_loss: 0.6593 - val_accuracy: 0.5500\n",
      "Epoch 5/12\n",
      "6/6 [==============================] - 0s 35ms/step - loss: 0.6697 - accuracy: 0.6538 - val_loss: 0.6392 - val_accuracy: 0.9000\n",
      "Epoch 6/12\n",
      "6/6 [==============================] - 0s 36ms/step - loss: 0.6559 - accuracy: 0.6538 - val_loss: 0.5964 - val_accuracy: 0.8000\n",
      "Epoch 7/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6369 - accuracy: 0.6603 - val_loss: 0.5356 - val_accuracy: 0.8500\n",
      "Epoch 8/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6256 - accuracy: 0.5962 - val_loss: 0.5018 - val_accuracy: 0.8000\n",
      "Epoch 9/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6236 - accuracy: 0.6090 - val_loss: 0.4819 - val_accuracy: 0.8750\n",
      "Epoch 10/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6428 - accuracy: 0.6538 - val_loss: 0.4949 - val_accuracy: 0.9000\n",
      "Epoch 11/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6141 - accuracy: 0.6410 - val_loss: 0.5402 - val_accuracy: 0.7750\n",
      "Epoch 12/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6240 - accuracy: 0.6410 - val_loss: 0.5286 - val_accuracy: 0.8500\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.History at 0x2fc0b4810>"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_gru.fit(x_train, y_train, epochs=12, batch_size=30, validation_data=(x_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2/2 [==============================] - 0s 7ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GRU Accuracy:  0.85\n"
     ]
    }
   ],
   "source": [
    "y_hat = model_gru.predict(x_test)\n",
    "y_hat = np.round(y_hat).flatten()\n",
    "gru_acc = accuracy_score(y_test, y_hat)\n",
    "\n",
    "cm = confusion_matrix(y_test, y_hat)\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=['Normal', 'Glaucoma'])\n",
    "disp.plot()\n",
    "plt.show()\n",
    "print(\"GRU Accuracy: \", gru_acc)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_lstm = Sequential()\n",
    "model_lstm.add(LSTM(62, return_sequences=True, input_shape=(62, 1)))\n",
    "model_lstm.add(Dropout(0.2))\n",
    "model_lstm.add(LSTM(50, return_sequences=True))\n",
    "model_lstm.add(Dropout(0.2))\n",
    "model_lstm.add(LSTM(50, return_sequences=False))\n",
    "model_lstm.add(Dropout(0.2))\n",
    "model_lstm.add(Dense(1, activation = \"sigmoid\"))\n",
    "\n",
    "model_lstm.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/12\n",
      "6/6 [==============================] - 3s 135ms/step - loss: 0.6954 - accuracy: 0.4423 - val_loss: 0.6897 - val_accuracy: 0.4750\n",
      "Epoch 2/12\n",
      "6/6 [==============================] - 0s 35ms/step - loss: 0.6906 - accuracy: 0.5064 - val_loss: 0.6852 - val_accuracy: 0.4750\n",
      "Epoch 3/12\n",
      "6/6 [==============================] - 0s 37ms/step - loss: 0.6855 - accuracy: 0.5064 - val_loss: 0.6800 - val_accuracy: 0.5000\n",
      "Epoch 4/12\n",
      "6/6 [==============================] - 0s 36ms/step - loss: 0.6807 - accuracy: 0.5064 - val_loss: 0.6773 - val_accuracy: 0.4750\n",
      "Epoch 5/12\n",
      "6/6 [==============================] - 0s 35ms/step - loss: 0.6882 - accuracy: 0.5128 - val_loss: 0.6739 - val_accuracy: 0.5000\n",
      "Epoch 6/12\n",
      "6/6 [==============================] - 0s 35ms/step - loss: 0.6803 - accuracy: 0.5449 - val_loss: 0.6602 - val_accuracy: 0.5250\n",
      "Epoch 7/12\n",
      "6/6 [==============================] - 0s 37ms/step - loss: 0.6733 - accuracy: 0.6218 - val_loss: 0.6531 - val_accuracy: 0.7250\n",
      "Epoch 8/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6692 - accuracy: 0.6603 - val_loss: 0.6361 - val_accuracy: 0.6500\n",
      "Epoch 9/12\n",
      "6/6 [==============================] - 0s 34ms/step - loss: 0.6604 - accuracy: 0.6346 - val_loss: 0.6230 - val_accuracy: 0.6500\n",
      "Epoch 10/12\n",
      "6/6 [==============================] - 0s 35ms/step - loss: 0.6506 - accuracy: 0.6410 - val_loss: 0.5954 - val_accuracy: 0.8500\n",
      "Epoch 11/12\n",
      "6/6 [==============================] - 0s 38ms/step - loss: 0.6269 - accuracy: 0.6603 - val_loss: 0.5619 - val_accuracy: 0.9000\n",
      "Epoch 12/12\n",
      "6/6 [==============================] - 0s 40ms/step - loss: 0.6124 - accuracy: 0.6731 - val_loss: 0.5856 - val_accuracy: 0.6250\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.src.callbacks.History at 0x2ebd9add0>"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_lstm.fit(x_train, y_train, epochs=12, batch_size=30, validation_data=(x_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2/2 [==============================] - 0s 8ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LSTM Accuracy:  0.625\n"
     ]
    }
   ],
   "source": [
    "y_hat = model_lstm.predict(x_test)\n",
    "y_hat = np.round(y_hat).flatten()\n",
    "lstm_acc = accuracy_score(y_test, y_hat)\n",
    "\n",
    "cm = confusion_matrix(y_test, y_hat)\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=['Normal', 'Glaucoma'])\n",
    "disp.plot()\n",
    "plt.show()\n",
    "print(\"LSTM Accuracy: \", lstm_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAAA5tklEQVR4nO3de1yUZf7/8fcAMuABUhHwQKCS5wMqQmZarSQWa5nW4ikV08o0demgWELmJmlm7KbJZp62lSRN235llo1rrUlLSmSup7QUTEHJBMUCg/n90bdpZ8HD4MDgzev5eNyPnGuu67o/tzUz7+77mntMVqvVKgAAAINwc3UBAAAAzkS4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4Aeq4b7/9VlOmTFG7du1Uv3591a9fX506ddLkyZO1e/duu77PPPOMTCaTbatXr55CQkI0depUnTlzpsLcJpNJU6ZMqXS/69evl8lk0rZt26rhqK7cr8dUUFBwyX5HjhxRXFyc2rZtKy8vLwUGBqp///5KSkqSJK1atcru7+ZiW0hIiN1+3dzclJubW2F/RUVF8vb2vuTfIYDKebi6AACu8+677yo2NlYeHh4aNWqUunfvLjc3N+3fv18bNmzQ0qVL9e233yo4ONhu3NKlS9WwYUMVFxfLYrHo5ZdfVlZWlrZv3+6iI6lehw4dUu/eveXt7a3x48crJCREJ06cUFZWlubPn685c+aof//+ev311+3GTZgwQREREXrwwQdtbQ0bNrTrYzab9cYbb+jJJ5+0a9+wYUP1HRBgcIQboI46fPiwhg8fruDgYFksFjVv3tzu+fnz5+uVV16Rm1vFE7z33nuv/Pz8JEkPPfSQhg8frvT0dGVmZioiIqJG6q9JL730ks6dO6fs7OwKQe/kyZOSpDZt2qhNmzZ2zz388MNq06aNRo8efdG577zzzkrDTVpammJiYvTWW2856SiAuoPLUkAdtWDBAhUXF2vlypUVgo0keXh4aOrUqQoKCrrsXP369ZP0S2Bypl8vXX388ccVnvvrX/8qk8mkPXv2SJLy8vIUFxenVq1ayWw2q3nz5rr77rt15MiRq67j8OHDatWqVYVgI0n+/v5XNffIkSOVnZ2t/fv329ry8vK0detWjRw58qrmBuoqwg1QR7377rsKDQ1VZGTkVc/1a4Bo3LjxVc/132JiYtSwYUO9+eabFZ5LT09X586d1aVLF0nSsGHDtHHjRsXFxemVV17R1KlTdfbsWeXk5Fx1HcHBwcrNzdXWrVuveq7/1b9/f7Vq1UppaWm2tvT0dDVs2FAxMTFO3x9QFxBugDqoqKhIx48ftwWD/3bmzBkVFBTYth9//LFCn9OnT6ugoEBHjx7VypUrtWTJEjVr1kz9+/d3ap3e3t4aPHiw1q9fr7KyMlt7Xl6ePv74Y8XGxtpq3rFjh55++mnNnTtXDzzwgBISErR161an1DR16lR5enpqwIAB6tGjh6ZPn65//OMfOn/+/FXPbTKZNHz4cL3xxhu2tjVr1mjo0KEym81XPT9QFxFugDqoqKhIUsXFrZJ06623qlmzZrZtyZIlFfq0b99ezZo1U0hIiMaPH6/Q0FC9//77ql+/vtNrjY2N1cmTJ+2+VbV+/XqVl5fbwo23t7c8PT21bds2/fDDD06voXPnzsrOztbo0aN15MgR/fnPf9aQIUMUEBCgZcuWXfX8I0eO1KFDh/T555/b/sklKaDqCDdAHdSoUSNJ0rlz5yo899e//lVbtmzR3//+94uOf+utt7RlyxalpaXpxhtv1MmTJ+Xt7V2lWkwm0yWfHzRokHx9fZWenm5rS09PV1hYmNq1ayfpl28czZ8/X++//74CAgLUv39/LViwQHl5eVWqqTLt2rXT66+/roKCAu3evVvz5s2Th4eHHnzwQX300UdXNXePHj3UoUMHpaWlac2aNQoMDNTvfvc7J1UO1D2EG6AO8vX1VfPmzW2Lcf9bZGSkoqKi1Ldv34uO79+/v6KiojRixAht2bJF3t7eGjVqlMrLy+36mc3mSi9rSbJd0vHy8rpkrWazWUOGDNHGjRv1888/67vvvtOnn35qO2vzq+nTp+vgwYNKTk6Wl5eXZs+erY4dO+qLL7645PyOcnd3V9euXZWQkKCNGzdK+uUy0tUaOXKk0tPTlZaWptjY2Eq/pQbgyvDqAeqomJgYHTp0SJmZmVc1T8OGDZWUlKTs7OwKC3+Dg4N14MCBSsf92l7ZN5D+V2xsrAoKCmSxWLRu3TpZrdYK4UaS2rZtq8cee0wffvih9uzZo9LSUr344otVOKorEx4eLkk6ceLEVc81cuRInThxQgcPHuSSFHCVCDdAHfXkk0+qfv36Gj9+vPLz8ys8b7Var3iuUaNGqVWrVpo/f75d+5133qnPPvtMu3btsms/c+aM1qxZo7CwMAUGBl52/qioKDVp0kTp6elKT09XRESEWrdubXv+/Pnz+umnn+zGtG3bVo0aNVJJSckVH8fF/Otf/9KFCxcqtG/atEnSL2uQrlbbtm2VkpKi5ORkQ94rCKhJ3MQPqKNuuOEGpaWlacSIEWrfvr3tDsVWq1Xffvut0tLS5ObmplatWl12rnr16mnatGl64okntHnzZg0aNEiSNHPmTK1bt079+/fXQw89pA4dOuj48eNatWqVTpw4oZUrV15RrfXq1dPQoUO1du1aFRcXa+HChXbPHzx4UAMGDNAf/vAHderUSR4eHtq4caPy8/M1fPjwK9rHokWLKiyIdnNz06xZszR//nzt2rVLQ4cOVbdu3SRJWVlZ+tvf/qYmTZpo+vTpV7SPy5k2bZpT5gHqPCuAOu3QoUPWSZMmWUNDQ61eXl5Wb29va4cOHawPP/ywNTs7265vUlKSVZL11KlTFeYpLCy0+vr6Wm+55Ra79mPHjlknTJhgbdmypdXDw8PapEkT6+9//3vrZ5995lCdW7ZssUqymkwma25urt1zBQUF1smTJ1s7dOhgbdCggdXX19caGRlpffPNNy8776/HVNnm7u5utVqt1k8//dQ6efJka5cuXay+vr7WevXqWa+//nrruHHjrIcPH77o3A0aNLCOHTv2kvut7O/yv0myTp48+bLHAeA3JqvVgXPPAAAAtRxrbgAAgKEQbgAAgKEQbgAAgKEQbgAAgKEQbgAAgKEQbgAAgKHUuZv4lZeX6/jx42rUqNFlf7APAADUDlarVWfPnlWLFi0u+9trdS7cHD9+XEFBQa4uAwAAVEFubu5l75xe58JNo0aNJP3yl+Pj4+PiagAAwJUoKipSUFCQ7XP8UupcuPn1UpSPjw/hBgCAa8yVLClhQTEAADAUwg0AADAUwg0AADAUwg0AADAUwg0AADAUwg0Mr6ysTLNnz1br1q3l7e2ttm3bau7cubJarbY+48aNk8lkstsGDRp02bmXLFmikJAQeXl5KTIyUpmZmdV5KACAK1DnvgqOumf+/PlaunSpVq9erc6dO2vnzp2Ki4uTr6+vpk6daus3aNAgrVy50vbYbDZfct709HTFx8crNTVVkZGRSklJUXR0tA4cOCB/f/9qOx4AwKVx5gaGt2PHDt19992KiYlRSEiI7r33Xg0cOLDCWRaz2azAwEDb1rhx40vOu2jRIk2cOFFxcXHq1KmTUlNTVb9+fa1YsaI6DwcAcBmEGxjeTTfdJIvFooMHD0qSvvzyS23fvl133HGHXb9t27bJ399f7du316RJk/T9999fdM7S0lLt2rVLUVFRtjY3NzdFRUUpIyOjeg4EAHBFuCwFw5s5c6aKiorUoUMHubu7q6ysTM8995xGjRpl6zNo0CANHTpUrVu31uHDhzVr1izdcccdysjIkLu7e4U5CwoKVFZWpoCAALv2gIAA7d+/v9qPCQBwcYQbGN6bb76pNWvWKC0tTZ07d1Z2dramT5+uFi1aaOzYsZKk4cOH2/p37dpV3bp1U9u2bbVt2zYNGDDAVaUDAKqAcAPDe+KJJzRz5kxbgOnatauOHj2q5ORkW7j5X23atJGfn58OHTpUabjx8/OTu7u78vPz7drz8/MVGBjo/IMAAFwx1tzA8M6fPy83N/v/1N3d3VVeXn7RMceOHdP333+v5s2bV/q8p6enevXqJYvFYmsrLy+XxWJRnz59nFM4AKBKCDcwvMGDB+u5557Te++9pyNHjmjjxo1atGiR7rnnHknSuXPn9MQTT+izzz7TkSNHZLFYdPfddys0NFTR0dG2eQYMGKDFixfbHsfHx2vZsmVavXq19u3bp0mTJqm4uFhxcXE1fowAgN9wWQqG9/LLL2v27Nl65JFHdPLkSbVo0UIPPfSQEhMTJf1yFmf37t1avXq1zpw5oxYtWmjgwIGaO3eu3b1uDh8+rIKCAtvj2NhYnTp1SomJicrLy1NYWJg2b95cYZExAKBmmaz/fZvWOqCoqEi+vr4qLCyUj4+P0+c3zTE5fU7AKKxJdertBoATOfL5zWUpAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKIQbAABgKC4PN0uWLFFISIi8vLwUGRmpzMzMS/ZPSUlR+/bt5e3traCgIP3xj3/UTz/9VEPVAgCA2s6l4SY9PV3x8fFKSkpSVlaWunfvrujoaJ08ebLS/mlpaZo5c6aSkpK0b98+LV++XOnp6Zo1a1YNVw4AAGorl4abRYsWaeLEiYqLi1OnTp2Umpqq+vXra8WKFZX237Fjh/r27auRI0cqJCREAwcO1IgRIy57tgcAANQdLgs3paWl2rVrl6Kion4rxs1NUVFRysjIqHTMTTfdpF27dtnCzDfffKNNmzbpzjvvvOh+SkpKVFRUZLcBAADj8nDVjgsKClRWVqaAgAC79oCAAO3fv7/SMSNHjlRBQYFuvvlmWa1W/fzzz3r44YcveVkqOTlZc+bMcWrtAACg9nL5gmJHbNu2TfPmzdMrr7yirKwsbdiwQe+9957mzp170TEJCQkqLCy0bbm5uTVYMQAAqGkuO3Pj5+cnd3d35efn27Xn5+crMDCw0jGzZ8/W/fffrwkTJkiSunbtquLiYj344IN66qmn5OZWMauZzWaZzWbnHwAAAKiVXHbmxtPTU7169ZLFYrG1lZeXy2KxqE+fPpWOOX/+fIUA4+7uLkmyWq3VVywAALhmuOzMjSTFx8dr7NixCg8PV0REhFJSUlRcXKy4uDhJ0pgxY9SyZUslJydLkgYPHqxFixapR48eioyM1KFDhzR79mwNHjzYFnIAAEDd5tJwExsbq1OnTikxMVF5eXkKCwvT5s2bbYuMc3Jy7M7UPP300zKZTHr66af13XffqVmzZho8eLCee+45Vx0CAACoZUzWOnY9p6ioSL6+viosLJSPj4/T5zfNMTl9TsAorEl16u0GgBM58vl9TX1bCgAA4HIINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFAINwAAwFBqRbhZsmSJQkJC5OXlpcjISGVmZl6076233iqTyVRhi4mJqcGKAQBAbeXycJOenq74+HglJSUpKytL3bt3V3R0tE6ePFlp/w0bNujEiRO2bc+ePXJ3d9d9991Xw5UDAIDayOXhZtGiRZo4caLi4uLUqVMnpaamqn79+lqxYkWl/Zs0aaLAwEDbtmXLFtWvX59wAwAAJLk43JSWlmrXrl2Kioqytbm5uSkqKkoZGRlXNMfy5cs1fPhwNWjQoNLnS0pKVFRUZLcBAADjcmm4KSgoUFlZmQICAuzaAwIClJeXd9nxmZmZ2rNnjyZMmHDRPsnJyfL19bVtQUFBV103AACovVx+WepqLF++XF27dlVERMRF+yQkJKiwsNC25ebm1mCFAACgpnm4cud+fn5yd3dXfn6+XXt+fr4CAwMvOba4uFhr167Vs88+e8l+ZrNZZrP5qmsFAADXBpeeufH09FSvXr1ksVhsbeXl5bJYLOrTp88lx65bt04lJSUaPXp0dZcJAACuIS49cyNJ8fHxGjt2rMLDwxUREaGUlBQVFxcrLi5OkjRmzBi1bNlSycnJduOWL1+uIUOGqGnTpq4oGwAA1FIuDzexsbE6deqUEhMTlZeXp7CwMG3evNm2yDgnJ0dubvYnmA4cOKDt27frww8/dEXJAACgFjNZrVarq4uoSUVFRfL19VVhYaF8fHycPr9pjsnpcwJGYU2qU283AJzIkc/va/rbUgAAAP+LcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAMIzvvvtOo0ePVtOmTeXt7a2uXbtq586dkqQLFy5oxowZ6tq1qxo0aKAWLVpozJgxOn78+CXnfOaZZ2Qymey2Dh061MThoIo8XF0AAADO8MMPP6hv37667bbb9P7776tZs2b6+uuv1bhxY0nS+fPnlZWVpdmzZ6t79+764YcfNG3aNN111122AHQxnTt31kcffWR77OHBx2dtxr8dAIAhzJ8/X0FBQVq5cqWtrXXr1rY/+/r6asuWLXZjFi9erIiICOXk5Oj666+/6NweHh4KDAx0ftGoFlyWAgAYwjvvvKPw8HDdd9998vf3V48ePbRs2bJLjiksLJTJZNJ11113yX5ff/21WrRooTZt2mjUqFHKyclxYuVwNsINAMAQvvnmGy1dulQ33HCDPvjgA02aNElTp07V6tWrK+3/008/acaMGRoxYoR8fHwuOm9kZKRWrVqlzZs3a+nSpfr222/Vr18/nT17troOBVfJZLVara4uoiYVFRXJ19dXhYWFl/yPuapMc0xOnxMwCmtSnXq7QQ3z9PRUeHi4duzYYWubOnWqPv/8c2VkZNj1vXDhgoYNG6Zjx45p27ZtDn0enDlzRsHBwVq0aJEeeOABp9WPS3Pk85szNwAAQ2jevLk6depk19axY8cKl5AuXLigP/zhDzp69Ki2bNni8P/oXnfddWrXrp0OHTp01TWjehBuAACG0LdvXx04cMCu7eDBgwoODrY9/jXYfP311/roo4/UtGlTh/dz7tw5HT58WM2bN7/qmlE9CDcAAEP44x//qM8++0zz5s3ToUOHlJaWpldffVWTJ0+W9Euwuffee7Vz506tWbNGZWVlysvLU15enkpLS23zDBgwQIsXL7Y9fvzxx/Xxxx/ryJEj2rFjh+655x65u7trxIgRNX6MuDJ8FRwAYAi9e/fWxo0blZCQoGeffVatW7dWSkqKRo0aJemXG/y98847kqSwsDC7sf/85z916623SpIOHz6sgoIC23PHjh3TiBEj9P3336tZs2a6+eab9dlnn6lZs2Y1clxwHAuKnYwFxcDFGWZBsYnXOXBJ1RAtWFAMAADqLMINAAAwFMINAAAwFMINAAAwFMINAAAwFJeHmyVLligkJEReXl6KjIxUZmbmJfufOXNGkydPVvPmzWU2m9WuXTtt2rSphqoFAAC1nUvvc5Oenq74+HilpqYqMjJSKSkpio6O1oEDB+Tv71+hf2lpqW6//Xb5+/tr/fr1atmypY4ePXrZX3MFAAB1h0vDzaJFizRx4kTFxcVJklJTU/Xee+9pxYoVmjlzZoX+K1as0OnTp7Vjxw7Vq1dPkhQSElKTJQMAgFrOZZelSktLtWvXLkVFRf1WjJuboqKiKvx666/eeecd9enTR5MnT1ZAQIC6dOmiefPmqays7KL7KSkpUVFRkd0GAACMy2XhpqCgQGVlZQoICLBrDwgIUF5eXqVjvvnmG61fv15lZWXatGmTZs+erRdffFF/+tOfLrqf5ORk+fr62ragoCCnHgcAAKhdXL6g2BHl5eXy9/fXq6++ql69eik2NlZPPfWUUlNTLzomISFBhYWFti03N7cGKwYAADXNZWtu/Pz85O7urvz8fLv2/Px8BQYGVjqmefPmqlevntzd3W1tHTt2tP2iq6enZ4UxZrNZZrPZucUDAIBay2Vnbjw9PdWrVy9ZLBZbW3l5uSwWi/r06VPpmL59++rQoUMqLy+3tR08eFDNmzevNNgAAIC6x6WXpeLj47Vs2TKtXr1a+/bt06RJk1RcXGz79tSYMWOUkJBg6z9p0iSdPn1a06ZN08GDB/Xee+9p3rx5mjx5sqsOAQAA1DIu/Sp4bGysTp06pcTEROXl5SksLEybN2+2LTLOycmRm9tv+SsoKEgffPCB/vjHP6pbt25q2bKlpk2bphkzZrjqEAAAQC1jslqtVlcXUZOKiork6+urwsJC+fj4OH1+0xyT0+cEjMKaZJC3GxOvc+CSqiFaOPL5fU19WwoAAOByCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQHA43ISEhevbZZ5WTk1Md9QAAAFwVh8PN9OnTtWHDBrVp00a333671q5dq5KSkuqoDQAAwGFVCjfZ2dnKzMxUx44d9eijj6p58+aaMmWKsrKyqlTEkiVLFBISIi8vL0VGRiozM/OifVetWiWTyWS3eXl5VWm/AADAeKq85qZnz576y1/+ouPHjyspKUmvvfaaevfurbCwMK1YsUJWq/WK5klPT1d8fLySkpKUlZWl7t27Kzo6WidPnrzoGB8fH504ccK2HT16tKqHAQAADKbK4ebChQt68803ddddd+mxxx5TeHi4XnvtNQ0bNkyzZs3SqFGjrmieRYsWaeLEiYqLi1OnTp2Umpqq+vXra8WKFRcdYzKZFBgYaNsCAgKqehgAAMBgPBwdkJWVpZUrV+qNN96Qm5ubxowZo5deekkdOnSw9bnnnnvUu3fvy85VWlqqXbt2KSEhwdbm5uamqKgoZWRkXHTcuXPnFBwcrPLycvXs2VPz5s1T586dHT0UAABgQA6Hm969e+v222/X0qVLNWTIENWrV69Cn9atW2v48OGXnaugoEBlZWUVzrwEBARo//79lY5p3769VqxYoW7duqmwsFALFy7UTTfdpP/85z9q1apVhf4lJSV2C56LioouWxcAALh2ORxuvvnmGwUHB1+yT4MGDbRy5coqF3Upffr0UZ8+fWyPb7rpJnXs2FF//etfNXfu3Ar9k5OTNWfOnGqpBQAA1D4Or7k5efKk/v3vf1do//e//62dO3c6NJefn5/c3d2Vn59v156fn6/AwMArmqNevXrq0aOHDh06VOnzCQkJKiwstG25ubkO1QgAAK4tDoebyZMnVxoQvvvuO02ePNmhuTw9PdWrVy9ZLBZbW3l5uSwWi93ZmUspKyvTV199pebNm1f6vNlslo+Pj90GAACMy+HLUnv37lXPnj0rtPfo0UN79+51uID4+HiNHTtW4eHhioiIUEpKioqLixUXFydJGjNmjFq2bKnk5GRJ0rPPPqsbb7xRoaGhOnPmjF544QUdPXpUEyZMcHjfAADAeBwON2azWfn5+WrTpo1d+4kTJ+Th4fB0io2N1alTp5SYmKi8vDyFhYVp8+bNtkXGOTk5cnP77QTTDz/8oIkTJyovL0+NGzdWr169tGPHDnXq1MnhfQMAAOMxWa/0bnv/Z8SIETpx4oT+8Y9/yNfXV5J05swZDRkyRP7+/nrzzTerpVBnKSoqkq+vrwoLC6vlEpVpjsnpcwJGYU1y6O2m9jLxOgcuybFocUUc+fx2+FTLwoUL1b9/fwUHB6tHjx6SpOzsbAUEBOj111+vWsUAAABO4nC4admypXbv3q01a9boyy+/lLe3t+Li4jRixIhK73kDAABQkxxfJKNf7mPz4IMPOrsWAACAq1alcCP98q2pnJwclZaW2rXfddddV10UAABAVVXpDsX33HOPvvrqK5lMJtuvf5v+b4FdWVmZcysEAABwgMM38Zs2bZpat26tkydPqn79+vrPf/6jTz75ROHh4dq2bVs1lAgAAHDlHD5zk5GRoa1bt8rPz09ubm5yc3PTzTffrOTkZE2dOlVffPFFddQJAABwRRw+c1NWVqZGjRpJ+uW3oY4fPy5JCg4O1oEDB5xbHQAAgIMcPnPTpUsXffnll2rdurUiIyO1YMECeXp66tVXX61w12IAAICa5nC4efrpp1VcXCzpl995+v3vf69+/fqpadOmSk9Pd3qBAAAAjnA43ERHR9v+HBoaqv379+v06dNq3Lix7RtTAAAAruLQmpsLFy7Iw8NDe/bssWtv0qQJwQYAANQKDoWbevXq6frrr+deNgAAoNZy+NtSTz31lGbNmqXTp09XRz0AAABXxeE1N4sXL9ahQ4fUokULBQcHq0GDBnbPZ2VlOa04AAAARzkcboYMGVINZQAAADiHw+EmKSmpOuoAAABwCofX3AAAANRmDp+5cXNzu+TXvvkmFQAAcCWHw83GjRvtHl+4cEFffPGFVq9erTlz5jitMAAAgKpwONzcfffdFdruvfdede7cWenp6XrggQecUhgAAEBVOG3NzY033iiLxeKs6QAAAKrEKeHmxx9/1F/+8he1bNnSGdMBAABUmcOXpf73BzKtVqvOnj2r+vXr6+9//7tTiwMAAHCUw+HmpZdesgs3bm5uatasmSIjI9W4cWOnFgcAAOAoh8PNuHHjqqEMAAAA53B4zc3KlSu1bt26Cu3r1q3T6tWrnVIUAABAVTkcbpKTk+Xn51eh3d/fX/PmzXNKUQAAAFXlcLjJyclR69atK7QHBwcrJyfHKUUBAABUlcPhxt/fX7t3767Q/uWXX6pp06ZOKQoAAKCqHA43I0aM0NSpU/XPf/5TZWVlKisr09atWzVt2jQNHz68OmoEAAC4Yg6Hm7lz5yoyMlIDBgyQt7e3vL29NXDgQP3ud7+r8pqbJUuWKCQkRF5eXoqMjFRmZuYVjVu7dq1MJpOGDBlSpf0CAADjcfir4J6enkpPT9ef/vQnZWdny9vbW127dlVwcHCVCkhPT1d8fLxSU1MVGRmplJQURUdH68CBA/L397/ouCNHjujxxx9Xv379qrRfAABgTCar1Wp1ZQGRkZHq3bu3Fi9eLEkqLy9XUFCQHn30Uc2cObPSMWVlZerfv7/Gjx+vf/3rXzpz5ozefvvtK9pfUVGRfH19VVhYKB8fH2cdho1pjunynYA6yprk0rcb5zHxOgcuqRqihSOf3w5flho2bJjmz59foX3BggW67777HJqrtLRUu3btUlRU1G8FubkpKipKGRkZFx337LPPyt/f/4p+gbykpERFRUV2GwAAMC6Hw80nn3yiO++8s0L7HXfcoU8++cShuQoKClRWVqaAgAC79oCAAOXl5VU6Zvv27Vq+fLmWLVt2RftITk6Wr6+vbQsKCnKoRgAAcG1xONycO3dOnp6eFdrr1atX7WdFzp49q/vvv1/Lli2r9EaClUlISFBhYaFty83NrdYaAQCAazm8oLhr165KT09XYmKiXfvatWvVqVMnh+by8/OTu7u78vPz7drz8/MVGBhYof/hw4d15MgRDR482NZWXl4uSfLw8NCBAwfUtm1buzFms1lms9mhugAAwLXL4XAze/ZsDR06VIcPH9bvfvc7SZLFYlFaWprWr1/v0Fyenp7q1auXLBaL7evc5eXlslgsmjJlSoX+HTp00FdffWXX9vTTT+vs2bP685//zCUnAADgeLgZPHiw3n77bc2bN0/r16+Xt7e3unfvrq1bt6pJkyYOFxAfH6+xY8cqPDxcERERSklJUXFxseLi4iRJY8aMUcuWLZWcnCwvLy916dLFbvx1110nSRXaAQBA3eRwuJGkmJgYxcTESPrlq1lvvPGGHn/8ce3atUtlZWUOzRUbG6tTp04pMTFReXl5CgsL0+bNm22LjHNycuTm5vDSIAAAUEdV+T43n3zyiZYvX6633npLLVq00NChQzVs2DD17t3b2TU6Ffe5AVyH+9wAdYSL73Pj0JmbvLw8rVq1SsuXL1dRUZH+8Ic/qKSkRG+//bbDi4kBAACqwxVf7xk8eLDat2+v3bt3KyUlRcePH9fLL79cnbUBAAA47IrP3Lz//vuaOnWqJk2apBtuuKE6awIAAKiyKz5zs337dp09e1a9evVSZGSkFi9erIKCguqsDQAAwGFXHG5uvPFGLVu2TCdOnNBDDz2ktWvXqkWLFiovL9eWLVt09uzZ6qwTAADgijj8HesGDRpo/Pjx2r59u7766is99thjev755+Xv76+77rqrOmoEAAC4Yld1A5n27dtrwYIFOnbsmN544w1n1QQAAFBlTrk7nru7u4YMGaJ33nnHGdMBAABUGbf+BQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhkK4AQAAhlIrws2SJUsUEhIiLy8vRUZGKjMz86J9N2zYoPDwcF133XVq0KCBwsLC9Prrr9dgtQAAoDZzebhJT09XfHy8kpKSlJWVpe7duys6OlonT56stH+TJk301FNPKSMjQ7t371ZcXJzi4uL0wQcf1HDlAACgNjJZrVarKwuIjIxU7969tXjxYklSeXm5goKC9Oijj2rmzJlXNEfPnj0VExOjuXPnXrZvUVGRfH19VVhYKB8fn6uqvTKmOSanzwkYhTXJpW83zmPidQ5cUjVEC0c+v1165qa0tFS7du1SVFSUrc3NzU1RUVHKyMi47Hir1SqLxaIDBw6of//+lfYpKSlRUVGR3QYAAIzLpeGmoKBAZWVlCggIsGsPCAhQXl7eRccVFhaqYcOG8vT0VExMjF5++WXdfvvtlfZNTk6Wr6+vbQsKCnLqMQAAgNrF5WtuqqJRo0bKzs7W559/rueee07x8fHatm1bpX0TEhJUWFho23Jzc2u2WAAAUKM8XLlzPz8/ubu7Kz8/3649Pz9fgYGBFx3n5uam0NBQSVJYWJj27dun5ORk3XrrrRX6ms1mmc1mp9YNAABqL5eeufH09FSvXr1ksVhsbeXl5bJYLOrTp88Vz1NeXq6SkpLqKBEAAFxjXHrmRpLi4+M1duxYhYeHKyIiQikpKSouLlZcXJwkacyYMWrZsqWSk5Ml/bKGJjw8XG3btlVJSYk2bdqk119/XUuXLnXlYQAAgFrC5eEmNjZWp06dUmJiovLy8hQWFqbNmzfbFhnn5OTIze23E0zFxcV65JFHdOzYMXl7e6tDhw76+9//rtjYWFcdAgAAqEVcfp+bmsZ9bgDX4T43QB1Rl+9zAwAA4GyEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCi1ItwsWbJEISEh8vLyUmRkpDIzMy/ad9myZerXr58aN26sxo0bKyoq6pL9AQBA3eLycJOenq74+HglJSUpKytL3bt3V3R0tE6ePFlp/23btmnEiBH65z//qYyMDAUFBWngwIH67rvvarhyAABQG5msVqvVlQVERkaqd+/eWrx4sSSpvLxcQUFBevTRRzVz5szLji8rK1Pjxo21ePFijRkz5rL9i4qK5Ovrq8LCQvn4+Fx1/f/LNMfk9DkBo7AmufTtxnlMvM6BS6qGaOHI57dLz9yUlpZq165dioqKsrW5ubkpKipKGRkZVzTH+fPndeHCBTVp0qTS50tKSlRUVGS3AQAA43JpuCkoKFBZWZkCAgLs2gMCApSXl3dFc8yYMUMtWrSwC0j/LTk5Wb6+vrYtKCjoqusGAAC1l8vX3FyN559/XmvXrtXGjRvl5eVVaZ+EhAQVFhbattzc3BquEgAA1CQPV+7cz89P7u7uys/Pt2vPz89XYGDgJccuXLhQzz//vD766CN169btov3MZrPMZrNT6gUAALWfS8/ceHp6qlevXrJYLLa28vJyWSwW9enT56LjFixYoLlz52rz5s0KDw+viVIBAMA1wqVnbiQpPj5eY8eOVXh4uCIiIpSSkqLi4mLFxcVJksaMGaOWLVsqOTlZkjR//nwlJiYqLS1NISEhtrU5DRs2VMOGDV12HAAAoHZwebiJjY3VqVOnlJiYqLy8PIWFhWnz5s22RcY5OTlyc/vtBNPSpUtVWlqqe++9126epKQkPfPMMzVZOgAAqIVcfp+bmsZ9bgDX4T43QB1Rl+9zAwAA4GyEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCiEGwAAYCguDzdLlixRSEiIvLy8FBkZqczMzIv2/c9//qNhw4YpJCREJpNJKSkpNVcoAAC4Jrg03KSnpys+Pl5JSUnKyspS9+7dFR0drZMnT1ba//z582rTpo2ef/55BQYG1nC1AADgWuDScLNo0SJNnDhRcXFx6tSpk1JTU1W/fn2tWLGi0v69e/fWCy+8oOHDh8tsNtdwtQAA4FrgsnBTWlqqXbt2KSoq6rdi3NwUFRWljIwMp+2npKRERUVFdhsAADAul4WbgoIClZWVKSAgwK49ICBAeXl5TttPcnKyfH19bVtQUJDT5gYAALWPyxcUV7eEhAQVFhbattzcXFeXBAAAqpGHq3bs5+cnd3d35efn27Xn5+c7dbGw2WxmfQ4AAHWIy87ceHp6qlevXrJYLLa28vJyWSwW9enTx1VlAQCAa5zLztxIUnx8vMaOHavw8HBFREQoJSVFxcXFiouLkySNGTNGLVu2VHJysqRfFiHv3bvX9ufvvvtO2dnZatiwoUJDQ112HAAAoPZwabiJjY3VqVOnlJiYqLy8PIWFhWnz5s22RcY5OTlyc/vt5NLx48fVo0cP2+OFCxdq4cKFuuWWW7Rt27aaLh8AANRCJqvVanV1ETWpqKhIvr6+KiwslI+Pj9PnN80xOX1OwCisSQZ5uzHxOgcuqRqihSOf34b/thQAAKhbCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQCDcAAMBQakW4WbJkiUJCQuTl5aXIyEhlZmZesv+6devUoUMHeXl5qWvXrtq0aVMNVQoAAGo7l4eb9PR0xcfHKykpSVlZWerevbuio6N18uTJSvvv2LFDI0aM0AMPPKAvvvhCQ4YM0ZAhQ7Rnz54arhwAANRGJqvVanVlAZGRkerdu7cWL14sSSovL1dQUJAeffRRzZw5s0L/2NhYFRcX691337W13XjjjQoLC1Nqaupl91dUVCRfX18VFhbKx8fHeQfyf0xzTE6fEzAKa5JL326cx8TrHLikaogWjnx+u/TMTWlpqXbt2qWoqChbm5ubm6KiopSRkVHpmIyMDLv+khQdHX3R/gAAoG7xcOXOCwoKVFZWpoCAALv2gIAA7d+/v9IxeXl5lfbPy8urtH9JSYlKSkpsjwsLCyX9kgCrxU/VMy1gBNX2ugNQu1TDa/3X948rueDk0nBTE5KTkzVnzpwK7UFBQS6oBqjbfJ/3dXUJAGqCb/W91s+ePSvfy8zv0nDj5+cnd3d35efn27Xn5+crMDCw0jGBgYEO9U9ISFB8fLztcXl5uU6fPq2mTZvKxHVzQysqKlJQUJByc3OrZX0VgNqB13rdYLVadfbsWbVo0eKyfV0abjw9PdWrVy9ZLBYNGTJE0i/hw2KxaMqUKZWO6dOnjywWi6ZPn25r27Jli/r06VNpf7PZLLPZbNd23XXXOaN8XCN8fHx4wwPqAF7rxne5Mza/cvllqfj4eI0dO1bh4eGKiIhQSkqKiouLFRcXJ0kaM2aMWrZsqeTkZEnStGnTdMstt+jFF19UTEyM1q5dq507d+rVV1915WEAAIBawuXhJjY2VqdOnVJiYqLy8vIUFhamzZs32xYN5+TkyM3tty913XTTTUpLS9PTTz+tWbNm6YYbbtDbb7+tLl26uOoQAABALeLy+9wA1aWkpETJyclKSEiocGkSgHHwWsf/ItwAAABDcfnPLwAAADgT4QYAABgK4QYAABgK4QYAABgK4QbXlLy8PE2bNk2hoaHy8vJSQECA+vbtq6VLl+r8+fOSpJCQEJlMJplMJtWvX19du3bVa6+9ZjfPqlWrLnozR5PJpLfffruajwSAJI0bN852E9f/9eWXX+quu+6Sv7+/vLy8FBISotjYWJ08eVLPPPOM7XV+se3X+U0mkx5++OEK80+ePFkmk0njxo2rxiOEKxBucM345ptv1KNHD3344YeaN2+evvjiC2VkZOjJJ5/Uu+++q48++sjW99lnn9WJEye0Z88ejR49WhMnTtT777/vwuoBOOLUqVMaMGCAmjRpog8++ED79u3TypUr1aJFCxUXF+vxxx/XiRMnbFurVq1sr/tft18FBQVp7dq1+vHHH21tP/30k9LS0nT99de74vBQzVx+Ez/gSj3yyCPy8PDQzp071aBBA1t7mzZtdPfdd9v9UmyjRo1svzc2Y8YMLViwQFu2bNEdd9xR43UDcNynn36qwsJCvfbaa/Lw+OWjqnXr1rrttttsfRo2bGj7s7u7u93r/r/17NlThw8f1oYNGzRq1ChJ0oYNG3T99derdevW1XwkcAXO3OCa8P333+vDDz/U5MmT7YLNf6vsh1DLy8v11ltv6YcffpCnp2d1lwnASQIDA/Xzzz9r48aNcsbt2MaPH6+VK1faHq9YscL2Mz8wHsINrgmHDh2S1WpV+/bt7dr9/PzUsGFDNWzYUDNmzLC1z5gxQw0bNpTZbNa9996rxo0ba8KECTVdNoAquvHGGzVr1iyNHDlSfn5+uuOOO/TCCy8oPz+/SvONHj1a27dv19GjR3X06FF9+umnGj16tJOrRm1BuME1LTMzU9nZ2ercubNKSkps7U888YSys7O1detWRUZG6qWXXlJoaKgLKwXgqOeee055eXlKTU1V586dlZqaqg4dOuirr75yeK5mzZopJiZGq1at0sqVKxUTEyM/P79qqBq1AeEG14TQ0FCZTCYdOHDArr1NmzYKDQ2Vt7e3Xbufn59CQ0PVr18/rVu3TlOnTtXevXttz/v4+Ki4uFjl5eV2486cOSNJ8vX1rZ4DAeCQpk2b6r777tPChQu1b98+tWjRQgsXLqzSXOPHj9eqVau0evVqjR8/3smVojYh3OCa0LRpU91+++1avHixiouLHRobFBSk2NhYJSQk2Nrat2+vn3/+WdnZ2XZ9s7KyJEnt2rW76poBOJenp6fatm3r8HvArwYNGqTS0lJduHBB0dHRTq4OtQnflsI145VXXlHfvn0VHh6uZ555Rt26dZObm5s+//xz7d+/X7169bro2GnTpqlLly7auXOnwsPD1blzZw0cOFDjx4/Xiy++qDZt2ujAgQOaPn26YmNj1bJlyxo8MqBuKywsrPA/Gl999ZU++OADDR8+XO3atZPVatX/+3//T5s2bbJbGOwId3d37du3z/ZnGBfhBteMtm3b6osvvtC8efOUkJCgY8eOyWw2q1OnTnr88cf1yCOPXHRsp06dNHDgQCUmJmrTpk2SpPT0dCUlJemhhx7S8ePH1apVK91zzz2aPXt2TR0SAEnbtm1Tjx497Npuu+02hYaG6rHHHlNubq7MZrNuuOEGvfbaa7r//vurvC8fH5+rLRfXAJPVGd+xAwAAqCVYcwMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAMAAAyFcAPA8LZt2yaTyWT77bArERISopSUlGqrCUD1IdwAcLlx48bJZDLp4YcfrvDc5MmTZTKZNG7cuJovDMA1iXADoFYICgrS2rVr9eOPP9rafvrpJ6Wlpen66693YWUArjWEGwC1Qs+ePRUUFKQNGzbY2jZs2KDrr7/e7neHSkpKNHXqVPn7+8vLy0s333yzPv/8c7u5Nm3apHbt2snb21u33Xabjhw5UmF/27dvV79+/eTt7a2goCBNnTq1yr82DaB2IdwAqDXGjx9v94vPK1asUFxcnF2fJ598Um+99ZZWr16trKwshYaGKjo6WqdPn5Yk5ebmaujQoRo8eLCys7M1YcIEzZw5026Ow4cPa9CgQRo2bJh2796t9PR0bd++XVOmTKn+gwRQ7Qg3AGqN0aNHa/v27Tp69KiOHj2qTz/9VKNHj7Y9X1xcrKVLl+qFF17QHXfcoU6dOmnZsmXy9vbW8uXLJUlLly5V27Zt9eKLL6p9+/YaNWpUhfU6ycnJGjVqlKZPn64bbrhBN910k/7yl7/ob3/7m3766aeaPGQA1cDD1QUAwK+aNWummJgYrVq1SlarVTExMfLz87M9f/jwYV24cEF9+/a1tdWrV08RERHat2+fJGnfvn2KjIy0m7dPnz52j7/88kvt3r1ba9assbVZrVaVl5fr22+/VceOHavj8ADUEMINgFpl/PjxtstDS5YsqZZ9nDt3Tg899JCmTp1a4TkWLwPXPsINgFpl0KBBKi0tlclkUnR0tN1zbdu2laenpz799FMFBwdLki5cuKDPP/9c06dPlyR17NhR77zzjt24zz77zO5xz549tXfvXoWGhlbfgQBwGdbcAKhV3N3dtW/fPu3du1fu7u52zzVo0ECTJk3SE088oc2bN2vv3r2aOHGizp8/rwceeECS9PDDD+vrr7/WE088oQMHDigtLU2rVq2ym2fGjBnasWOHpkyZouzsbH399df6xz/+wYJiwCAINwBqHR8fH/n4+FT63PPPP69hw4bp/vvvV8+ePXXo0CF98MEHaty4saRfLiu99dZbevvtt9W9e3elpqZq3rx5dnN069ZNH3/8sQ4ePKh+/fqpR48eSkxMVIsWLar92ABUP5PVarW6uggAAABn4cwNAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwFMINAAAwlP8PymshPrVQ3jgAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(['GRU', 'LSTM'], [gru_acc, lstm_acc], color=['green', 'red'])\n",
    "plt.text(0, gru_acc, str(round(gru_acc*100, 3)))\n",
    "plt.text(1, lstm_acc, str(round(lstm_acc*100, 3)))\n",
    "plt.title('GRU vs LSTM')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Model')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}