deep_learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# HEAVILY DRAWN FROM: \n",
    "# https://www.tensorflow.org/tutorials/load_data/pandas_dataframe\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "BATCH_SIZE = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num GPUs Available:  0\n"
     ]
    }
   ],
   "source": [
    "print(\"Num GPUs Available: \", len(tf.config.list_physical_devices('GPU:0')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {},
   "outputs": [],
   "source": [
    "training_path = os.path.join(\"./\", \"final_results/\", \"trainment.csv\")\n",
    "test_set = pd.read_csv(\"./final_results/test.csv\")\n",
    "df = pd.read_csv(training_path)\n",
    "df = pd.concat([test_set, df])\n",
    "df = df.dropna(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [],
   "source": [
    "target = df.pop('target')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "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>polarity_title</th>\n",
       "      <th>subjectivity_title</th>\n",
       "      <th>polarity_description</th>\n",
       "      <th>subjectivity_description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.10</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>-0.10</td>\n",
       "      <td>0.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.10</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>-0.10</td>\n",
       "      <td>0.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.05</td>\n",
       "      <td>0.150000</td>\n",
       "      <td>-0.05</td>\n",
       "      <td>0.150000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.10</td>\n",
       "      <td>0.633333</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.633333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.25</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>-0.25</td>\n",
       "      <td>0.250000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   polarity_title  subjectivity_title  polarity_description  \\\n",
       "0           -0.10            0.200000                 -0.10   \n",
       "1           -0.10            0.200000                 -0.10   \n",
       "2           -0.05            0.150000                 -0.05   \n",
       "3            0.10            0.633333                  0.10   \n",
       "4           -0.25            0.250000                 -0.25   \n",
       "\n",
       "   subjectivity_description  \n",
       "0                  0.200000  \n",
       "1                  0.200000  \n",
       "2                  0.150000  \n",
       "3                  0.633333  \n",
       "4                  0.250000  "
      ]
     },
     "execution_count": 234,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "numeric_feature_names = [\n",
    "    'polarity_title', 'subjectivity_title',\n",
    "    'polarity_description', 'subjectivity_description', \n",
    "]\n",
    "numeric_features = df[numeric_feature_names]\n",
    "numeric_features.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(859, 4), dtype=float64, numpy=\n",
       "array([[-0.1       ,  0.2       , -0.1       ,  0.2       ],\n",
       "       [-0.1       ,  0.2       , -0.1       ,  0.2       ],\n",
       "       [-0.05      ,  0.15      , -0.05      ,  0.15      ],\n",
       "       ...,\n",
       "       [ 0.09444444,  0.15555556,  0.17327911,  0.4394066 ],\n",
       "       [ 0.        ,  0.        ,  0.30833333,  0.55833333],\n",
       "       [ 0.33214286,  0.70357143,  0.09209201,  0.3414066 ]])>"
      ]
     },
     "execution_count": 235,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.convert_to_tensor(numeric_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [],
   "source": [
    "normalizer = tf.keras.layers.Normalization(axis=-1)\n",
    "normalizer.adapt(numeric_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "430/430 [==============================] - 4s 7ms/step - loss: 0.6760 - accuracy: 0.6112\n",
      "Epoch 2/15\n",
      "430/430 [==============================] - 4s 9ms/step - loss: 0.6616 - accuracy: 0.6135\n",
      "Epoch 3/15\n",
      "430/430 [==============================] - 3s 7ms/step - loss: 0.6584 - accuracy: 0.6135\n",
      "Epoch 4/15\n",
      "430/430 [==============================] - 2s 6ms/step - loss: 0.6560 - accuracy: 0.6135\n",
      "Epoch 5/15\n",
      "430/430 [==============================] - 4s 9ms/step - loss: 0.6535 - accuracy: 0.6135\n",
      "Epoch 6/15\n",
      "430/430 [==============================] - 3s 7ms/step - loss: 0.6514 - accuracy: 0.6135\n",
      "Epoch 7/15\n",
      "430/430 [==============================] - 3s 7ms/step - loss: 0.6514 - accuracy: 0.6170\n",
      "Epoch 8/15\n",
      "430/430 [==============================] - 3s 7ms/step - loss: 0.6498 - accuracy: 0.6158\n",
      "Epoch 9/15\n",
      "430/430 [==============================] - 2s 5ms/step - loss: 0.6492 - accuracy: 0.6193\n",
      "Epoch 10/15\n",
      "430/430 [==============================] - 3s 8ms/step - loss: 0.6478 - accuracy: 0.6193\n",
      "Epoch 11/15\n",
      "430/430 [==============================] - 3s 6ms/step - loss: 0.6475 - accuracy: 0.6193\n",
      "Epoch 12/15\n",
      "430/430 [==============================] - 3s 7ms/step - loss: 0.6468 - accuracy: 0.6193\n",
      "Epoch 13/15\n",
      "430/430 [==============================] - 3s 6ms/step - loss: 0.6453 - accuracy: 0.6205\n",
      "Epoch 14/15\n",
      "430/430 [==============================] - 3s 8ms/step - loss: 0.6455 - accuracy: 0.6182\n",
      "Epoch 15/15\n",
      "430/430 [==============================] - 3s 7ms/step - loss: 0.6444 - accuracy: 0.6193\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f88daa3d790>"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_basic_model():\n",
    "  model = tf.keras.Sequential([\n",
    "    normalizer,\n",
    "    tf.keras.layers.Dense(10, activation='relu'),\n",
    "    tf.keras.layers.Dense(10, activation='relu'),\n",
    "    tf.keras.layers.Dense(1)\n",
    "  ])\n",
    "\n",
    "  model.compile(optimizer='adam',\n",
    "                loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),\n",
    "                metrics=['accuracy'])\n",
    "  return model\n",
    "\n",
    "model = get_basic_model()\n",
    "model.fit(numeric_features, target, epochs=15, batch_size=BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-05-23 10:14:46.200322: I tensorflow/core/common_runtime/process_util.cc:146] Creating new thread pool with default inter op setting: 2. Tune using inter_op_parallelism_threads for best performance.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_23\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " normalization_13 (Normaliza  (None, 4)                9         \n",
      " tion)                                                           \n",
      "                                                                 \n",
      " dense_39 (Dense)            (None, 10)                50        \n",
      "                                                                 \n",
      " dense_40 (Dense)            (None, 10)                110       \n",
      "                                                                 \n",
      " dense_41 (Dense)            (None, 1)                 11        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 180\n",
      "Trainable params: 171\n",
      "Non-trainable params: 9\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.models.load_model('saved_model/my_model')\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _update_step_xla while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: saved_model/my_model/assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: saved_model/my_model/assets\n"
     ]
    }
   ],
   "source": [
    "model.save('saved_model/my_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_set = pd.read_csv(\"./final_results/test.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13/13 - 0s - loss: 0.6484 - accuracy: 0.6218 - 232ms/epoch - 18ms/step\n",
      "\n",
      "Test accuracy: 0.6218274235725403\n"
     ]
    }
   ],
   "source": [
    "numeric_feature_names = [\n",
    "    'polarity_title', 'subjectivity_title',\n",
    "    'polarity_description', 'subjectivity_description', \n",
    "]\n",
    "numeric_features = test_set[numeric_feature_names]\n",
    "tf.convert_to_tensor(numeric_features)\n",
    "target = test_set.pop('target')\n",
    "test_loss, test_acc = model.evaluate(x = numeric_features, y = target,verbose=2)\n",
    "print('\\nTest accuracy:', test_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13/13 [==============================] - 0s 8ms/step\n"
     ]
    }
   ],
   "source": [
    "probability_model = tf.keras.Sequential([model, \n",
    "                                         tf.keras.layers.Softmax()])\n",
    "predictions = probability_model.predict(numeric_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGxCAYAAABvIsx7AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy88F64QAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA31ElEQVR4nO3de1xVVcL/8e9RBLxxMlDA5AHNUrwrFoJjahlqN516EkfDbBQvUxOmZTlWXuZp0KbMtDTtScwmLz2jPlmZiqZm3lIGNJPQHBHHIC8paI6gsn5/+PM8HgEV9QgsP+/Xa79e7r3XXnvts1nwde919nYYY4wAAAAsUqmsGwAAAHC9EXAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANbxKusGlIXCwkL99NNPqlmzphwOR1k3BwAAXAFjjI4fP666deuqUqVLX6O5KQPOTz/9pJCQkLJuBgAAuAr79+9XvXr1Llnmpgw4NWvWlHTuA/Lz8yvj1gAAgCuRl5enkJAQ19/xS7kpA87521J+fn4EHAAAKpgrGV7CIGMAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYJ0bEnCmTZum+vXry9fXVxEREVq3bl2JZRctWqT7779ftWvXlp+fn6KiorR8+fIi5RYuXKgmTZrIx8dHTZo00eLFiz15CAAAoALxeMBZsGCBhg0bptGjRys1NVUdOnRQ9+7dlZWVVWz5r7/+Wvfff7+WLl2qlJQUde7cWQ8//LBSU1NdZTZu3KjY2FjFxcVp27ZtiouLU69evbR582ZPHw4AAKgAHMYY48kdREZGqk2bNpo+fbprWXh4uHr27KnExMQrqqNp06aKjY3Vq6++KkmKjY1VXl6evvzyS1eZbt26qVatWpo3b95l68vLy5PT6VRubi7vogIAoIIozd9vj17BKSgoUEpKimJiYtyWx8TEaMOGDVdUR2FhoY4fP65bb73VtWzjxo1F6uzatesV1wkAAOzm0beJHz58WGfPnlVgYKDb8sDAQOXk5FxRHW+++aZ+/fVX9erVy7UsJyenVHXm5+crPz/fNZ+Xl3elhwAAACogjwac8y5+rbkx5opedT5v3jyNHTtWn376qerUqXPVdSYmJmrcuHGlbDVQVNhLX5R1E25KmRMeLOsmoIKiz5adsu63Hr1FFRAQoMqVKxe5snLw4MEiV2AutmDBAg0YMECffPKJunTp4rYuKCioVHWOGjVKubm5rmn//v1XcTQAAKCi8GjA8fb2VkREhJKTk92WJycnKzo6usTt5s2bp/79+2vu3Ll68MGiCTAqKqpInStWrCixTh8fH/n5+blNAADAXh6/RTV8+HDFxcWpbdu2ioqK0syZM5WVlaUhQ4ZIOnd15cCBA5ozZ46kc+GmX79+evvtt9WuXTvXlZqqVavK6XRKkhISEnTPPfdo4sSJ6tGjhz799FOtXLlS33zzjacPBwAAVAAefw5ObGysJk+erPHjx6tVq1b6+uuvtXTpUoWGhkqSsrOz3Z6JM2PGDJ05c0ZPP/20goODXVNCQoKrTHR0tObPn6+kpCS1aNFCs2fP1oIFCxQZGenpwwEAABWAx5+DUx7xHBxcLQYslo2yHqyIios+W3Y80W/LzXNwAAAAygIBBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOvckIAzbdo01a9fX76+voqIiNC6detKLJudna0+ffqoUaNGqlSpkoYNG1akzOzZs+VwOIpMp06d8uBRAACAisLjAWfBggUaNmyYRo8erdTUVHXo0EHdu3dXVlZWseXz8/NVu3ZtjR49Wi1btiyxXj8/P2VnZ7tNvr6+njoMAABQgXg84EyaNEkDBgzQwIEDFR4ersmTJyskJETTp08vtnxYWJjefvtt9evXT06ns8R6HQ6HgoKC3CYAAADJwwGnoKBAKSkpiomJcVseExOjDRs2XFPdJ06cUGhoqOrVq6eHHnpIqamp11QfAACwh0cDzuHDh3X27FkFBga6LQ8MDFROTs5V19u4cWPNnj1bS5Ys0bx58+Tr66v27dtr9+7dxZbPz89XXl6e2wQAAOx1QwYZOxwOt3ljTJFlpdGuXTs98cQTatmypTp06KBPPvlEd955p6ZOnVps+cTERDmdTtcUEhJy1fsGAADln0cDTkBAgCpXrlzkas3BgweLXNW5FpUqVdJdd91V4hWcUaNGKTc31zXt37//uu0bAACUPx4NON7e3oqIiFBycrLb8uTkZEVHR1+3/RhjlJaWpuDg4GLX+/j4yM/Pz20CAAD28vL0DoYPH664uDi1bdtWUVFRmjlzprKysjRkyBBJ566uHDhwQHPmzHFtk5aWJuncQOJDhw4pLS1N3t7eatKkiSRp3Lhxateune644w7l5eVpypQpSktL07vvvuvpwwEAABWAxwNObGysjhw5ovHjxys7O1vNmjXT0qVLFRoaKuncg/0ufiZO69atXf9OSUnR3LlzFRoaqszMTEnSsWPHNGjQIOXk5MjpdKp169b6+uuvdffdd3v6cAAAQAXgMMaYsm7EjZaXlyen06nc3FxuV6FUwl76oqybcFPKnPBgWTcBFRR9tux4ot+W5u8376ICAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgnRsScKZNm6b69evL19dXERERWrduXYlls7Oz1adPHzVq1EiVKlXSsGHDii23cOFCNWnSRD4+PmrSpIkWL17sodYDAICKxuMBZ8GCBRo2bJhGjx6t1NRUdejQQd27d1dWVlax5fPz81W7dm2NHj1aLVu2LLbMxo0bFRsbq7i4OG3btk1xcXHq1auXNm/e7MlDAQAAFYTDGGM8uYPIyEi1adNG06dPdy0LDw9Xz549lZiYeMltO3XqpFatWmny5Mluy2NjY5WXl6cvv/zStaxbt26qVauW5s2bd9k25eXlyel0Kjc3V35+fqU7INzUwl76oqybcFPKnPBgWTcBFRR9tux4ot+W5u+3R6/gFBQUKCUlRTExMW7LY2JitGHDhquud+PGjUXq7Nq16zXVCQAA7OHlycoPHz6ss2fPKjAw0G15YGCgcnJyrrrenJycUtWZn5+v/Px813xeXt5V7xsAAJR/N2SQscPhcJs3xhRZ5sk6ExMT5XQ6XVNISMg17RsAAJRvHg04AQEBqly5cpErKwcPHixyBaY0goKCSlXnqFGjlJub65r2799/1fsGAADln0cDjre3tyIiIpScnOy2PDk5WdHR0Vddb1RUVJE6V6xYUWKdPj4+8vPzc5sAAIC9PDoGR5KGDx+uuLg4tW3bVlFRUZo5c6aysrI0ZMgQSeeurhw4cEBz5sxxbZOWliZJOnHihA4dOqS0tDR5e3urSZMmkqSEhATdc889mjhxonr06KFPP/1UK1eu1DfffOPpwwEAABWAxwNObGysjhw5ovHjxys7O1vNmjXT0qVLFRoaKuncg/0ufiZO69atXf9OSUnR3LlzFRoaqszMTElSdHS05s+fr5dfflmvvPKKbr/9di1YsECRkZGePhwAAFABePw5OOURz8HB1eKZGmWD5+DgatFny47Vz8EBAAAoCwQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArHNDAs60adNUv359+fr6KiIiQuvWrbtk+bVr1yoiIkK+vr5q0KCB3nvvPbf1s2fPlsPhKDKdOnXKk4cBAAAqCI8HnAULFmjYsGEaPXq0UlNT1aFDB3Xv3l1ZWVnFlt+7d68eeOABdejQQampqfrTn/6kZ599VgsXLnQr5+fnp+zsbLfJ19fX04cDAAAqAC9P72DSpEkaMGCABg4cKEmaPHmyli9frunTpysxMbFI+ffee0//8R//ocmTJ0uSwsPDtXXrVr3xxht67LHHXOUcDoeCgoI83XwAAFABefQKTkFBgVJSUhQTE+O2PCYmRhs2bCh2m40bNxYp37VrV23dulWnT592LTtx4oRCQ0NVr149PfTQQ0pNTb3+BwAAACokjwacw4cP6+zZswoMDHRbHhgYqJycnGK3ycnJKbb8mTNndPjwYUlS48aNNXv2bC1ZskTz5s2Tr6+v2rdvr927dxdbZ35+vvLy8twmAABgrxsyyNjhcLjNG2OKLLtc+QuXt2vXTk888YRatmypDh066JNPPtGdd96pqVOnFltfYmKinE6nawoJCbmWwwEAAOWcRwNOQECAKleuXORqzcGDB4tcpTkvKCio2PJeXl7y9/cvdptKlSrprrvuKvEKzqhRo5Sbm+ua9u/ffxVHAwAAKgqPBhxvb29FREQoOTnZbXlycrKio6OL3SYqKqpI+RUrVqht27aqUqVKsdsYY5SWlqbg4OBi1/v4+MjPz89tAgAA9vL4Larhw4frv//7vzVr1iylp6frueeeU1ZWloYMGSLp3NWVfv36ucoPGTJE+/bt0/Dhw5Wenq5Zs2bpgw8+0PPPP+8qM27cOC1fvlz//Oc/lZaWpgEDBigtLc1VJwAAuLl5/GvisbGxOnLkiMaPH6/s7Gw1a9ZMS5cuVWhoqCQpOzvb7Zk49evX19KlS/Xcc8/p3XffVd26dTVlyhS3r4gfO3ZMgwYNUk5OjpxOp1q3bq2vv/5ad999t6cPBwAAVAAOc34E700kLy9PTqdTubm53K5CqYS99EVZN+GmlDnhwbJuAioo+mzZ8US/Lc3fb95FBQAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDo3JOBMmzZN9evXl6+vryIiIrRu3bpLll+7dq0iIiLk6+urBg0a6L333itSZuHChWrSpIl8fHzUpEkTLV682FPNBwAAFYzHA86CBQs0bNgwjR49WqmpqerQoYO6d++urKysYsvv3btXDzzwgDp06KDU1FT96U9/0rPPPquFCxe6ymzcuFGxsbGKi4vTtm3bFBcXp169emnz5s2ePhwAAFABOIwxxpM7iIyMVJs2bTR9+nTXsvDwcPXs2VOJiYlFyr/44otasmSJ0tPTXcuGDBmibdu2aePGjZKk2NhY5eXl6csvv3SV6datm2rVqqV58+Zdtk15eXlyOp3Kzc2Vn5/ftRwebjJhL31R1k24KWVOeLCsm4AKij5bdjzRb0vz99ujV3AKCgqUkpKimJgYt+UxMTHasGFDsdts3LixSPmuXbtq69atOn369CXLlFQnAAC4uXh5svLDhw/r7NmzCgwMdFseGBionJycYrfJyckptvyZM2d0+PBhBQcHl1impDrz8/OVn5/vms/Ly7uawwEAABWERwPOeQ6Hw23eGFNk2eXKX7y8NHUmJiZq3LhxpWrzteCSaNnx9K0MbpXYiT5bduiz8BSP3qIKCAhQ5cqVi1xZOXjwYJErMOcFBQUVW97Ly0v+/v6XLFNSnaNGjVJubq5r2r9//9UeEgAAqAA8GnC8vb0VERGh5ORkt+XJycmKjo4udpuoqKgi5VesWKG2bduqSpUqlyxTUp0+Pj7y8/NzmwAAgL08fotq+PDhiouLU9u2bRUVFaWZM2cqKytLQ4YMkXTu6sqBAwc0Z84cSee+MfXOO+9o+PDhio+P18aNG/XBBx+4fTsqISFB99xzjyZOnKgePXro008/1cqVK/XNN994+nAAAEAF4PGAExsbqyNHjmj8+PHKzs5Ws2bNtHTpUoWGhkqSsrOz3Z6JU79+fS1dulTPPfec3n33XdWtW1dTpkzRY4895ioTHR2t+fPn6+WXX9Yrr7yi22+/XQsWLFBkZKSnDwcAAFQAHn8OTnnk6efgMGCx7DCgEFeDPlt26LMojXLzHBwAAICyQMABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOh4NOEePHlVcXJycTqecTqfi4uJ07NixS25jjNHYsWNVt25dVa1aVZ06ddL333/vVqZTp05yOBxuU+/evT14JAAAoCLxaMDp06eP0tLStGzZMi1btkxpaWmKi4u75Davv/66Jk2apHfeeUdbtmxRUFCQ7r//fh0/ftytXHx8vLKzs13TjBkzPHkoAACgAvHyVMXp6elatmyZNm3apMjISEnS+++/r6ioKGVkZKhRo0ZFtjHGaPLkyRo9erQeffRRSdKHH36owMBAzZ07V4MHD3aVrVatmoKCgjzVfAAAUIF57ArOxo0b5XQ6XeFGktq1ayen06kNGzYUu83evXuVk5OjmJgY1zIfHx917NixyDYff/yxAgIC1LRpUz3//PNFrvAAAICbl8eu4OTk5KhOnTpFltepU0c5OTklbiNJgYGBbssDAwO1b98+13zfvn1Vv359BQUFaceOHRo1apS2bdum5OTkYuvNz89Xfn6+az4vL6/UxwMAACqOUl/BGTt2bJEBvhdPW7dulSQ5HI4i2xtjil1+oYvXX7xNfHy8unTpombNmql37976+9//rpUrV+of//hHsfUlJia6Bjo7nU6FhISU9rABAEAFUuorOM8888xlv7EUFham7du36+effy6y7tChQ0Wu0Jx3fkxNTk6OgoODXcsPHjxY4jaS1KZNG1WpUkW7d+9WmzZtiqwfNWqUhg8f7prPy8sj5AAAYLFSB5yAgAAFBARctlxUVJRyc3P17bff6u6775Ykbd68Wbm5uYqOji52m/O3nZKTk9W6dWtJUkFBgdauXauJEyeWuK/vv/9ep0+fdgtFF/Lx8ZGPj89l2wwAAOzgsUHG4eHh6tatm+Lj47Vp0yZt2rRJ8fHxeuihh9y+QdW4cWMtXrxY0rlbU8OGDdNf/vIXLV68WDt27FD//v1VrVo19enTR5K0Z88ejR8/Xlu3blVmZqaWLl2qxx9/XK1bt1b79u09dTgAAKAC8dggY+ncN52effZZ17eiHnnkEb3zzjtuZTIyMpSbm+uaHzlypP7973/rD3/4g44eParIyEitWLFCNWvWlCR5e3tr1apVevvtt3XixAmFhITowQcf1JgxY1S5cmVPHg4AAKggHMYYU9aNuNHy8vLkdDqVm5srPz+/615/2EtfXPc6cWUyJzxY1k1ABUSfLTv0WZRGaf5+8y4qAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1iHgAAAA6xBwAACAdQg4AADAOgQcAABgHQIOAACwDgEHAABYh4ADAACsQ8ABAADWIeAAAADrEHAAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWIeAAwAArEPAAQAA1vFowDl69Kji4uLkdDrldDoVFxenY8eOXXKbRYsWqWvXrgoICJDD4VBaWlqRMvn5+frjH/+ogIAAVa9eXY888oj+9a9/eeYgAABAhePRgNOnTx+lpaVp2bJlWrZsmdLS0hQXF3fJbX799Ve1b99eEyZMKLHMsGHDtHjxYs2fP1/ffPONTpw4oYceekhnz5693ocAAAAqIC9PVZyenq5ly5Zp06ZNioyMlCS9//77ioqKUkZGhho1alTsducDUGZmZrHrc3Nz9cEHH+ijjz5Sly5dJEl/+9vfFBISopUrV6pr167X/2AAAECF4rErOBs3bpTT6XSFG0lq166dnE6nNmzYcNX1pqSk6PTp04qJiXEtq1u3rpo1a3ZN9QIAAHt47ApOTk6O6tSpU2R5nTp1lJOTc031ent7q1atWm7LAwMDS6w3Pz9f+fn5rvm8vLyr3j8AACj/Sh1wxo4dq3Hjxl2yzJYtWyRJDoejyDpjTLHLr9Wl6k1MTLxsm6+nzAkP3rB9Abh29FnAPqUOOM8884x69+59yTJhYWHavn27fv755yLrDh06pMDAwNLu1iUoKEgFBQU6evSo21WcgwcPKjo6uthtRo0apeHDh7vm8/LyFBISctVtAAAA5VupA05AQIACAgIuWy4qKkq5ubn69ttvdffdd0uSNm/erNzc3BKDyJWIiIhQlSpVlJycrF69ekmSsrOztWPHDr3++uvFbuPj4yMfH5+r3icAAKhYPDbIODw8XN26dVN8fLw2bdqkTZs2KT4+Xg899JDbN6gaN26sxYsXu+Z/+eUXpaWlaefOnZKkjIwMpaWlucbXOJ1ODRgwQCNGjNCqVauUmpqqJ554Qs2bN3d9qwoAANzcPPocnI8//ljNmzdXTEyMYmJi1KJFC3300UduZTIyMpSbm+uaX7JkiVq3bq0HHzx3T7x3795q3bq13nvvPVeZt956Sz179lSvXr3Uvn17VatWTZ999pkqV67sycMBAAAVhMMYY8q6ETdaXl6enE6ncnNz5efnV9bNAQAAV6A0f795FxUAALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHUIOAAAwDoEHAAAYB0CDgAAsA4BBwAAWKfUbxO3wfm3U+Tl5ZVxSwAAwJU6/3f7St4ydVMGnOPHj0uSQkJCyrglAACgtI4fPy6n03nJMjflyzYLCwv1008/qWbNmnI4HGXdnHIjLy9PISEh2r9/Py8htQzn1l6cWztxXotnjNHx48dVt25dVap06VE2N+UVnEqVKqlevXpl3Yxyy8/Pjw5lKc6tvTi3duK8FnW5KzfnMcgYAABYh4ADAACsQ8CBi4+Pj8aMGSMfH5+ybgquM86tvTi3duK8XrubcpAxAACwG1dwAACAdQg4AADAOgQcC82ePVu33HLLNdeTmZkph8OhtLS0a64L/ycsLEyTJ0++5jLXqrQ/J2vWrJHD4dCxY8c81qabTUXvq2PHjlWrVq1u6D5vNPpr6fTv3189e/a8ofssCQEHJQoJCVF2draaNWsmyTMd5mb4BXk1tmzZokGDBl23+or7BRwbG6tdu3ZdcR3R0dHKzs52PYPiev1xvlB5+uVYkdyIvoqS3az9tby7KR/0h8srKCiQt7e3goKCyropN6XatWt7fB9Vq1ZV1apVr7g8Pw/lE3217NFfyyeu4JRDnTp10jPPPKNnnnlGt9xyi/z9/fXyyy+7Xi529OhR9evXT7Vq1VK1atXUvXt37d69u8T69uzZox49eigwMFA1atTQXXfdpZUrV7qVCQsL03/913+pf//+cjqdio+Pd7vsnZmZqc6dO0uSatWqJYfDof79+2vOnDny9/dXfn6+W32PPfaY+vXrd8njnD17tsaNG6dt27bJ4XDI4XBo9uzZV/GJ3Xh///vf1bx5c1WtWlX+/v7q0qWLfv31V3Xq1EnDhg1zK9uzZ0/179/fbdnx48fVp08f1ahRQ3Xr1tXUqVPd1l/8P7jc3FwNGjRIderUkZ+fn+69915t27bNbZslS5aobdu28vX1VUBAgB599FFJ536e9u3bp+eee871OUvu/6PLyMiQw+HQDz/84FbnpEmTFBYWJmOM21WBNWvW6KmnnlJubq6rzrFjx2r8+PFq3rx5kc8rIiJCr7766iU/07Fjx+rDDz/Up59+6qpzzZo1l9ymrN0sffW8CRMmKDAwUDVr1tSAAQN06tSp0nxcZYb+ev37qySdPXtWw4cPd/3sjxw58opegnnDGJQ7HTt2NDVq1DAJCQnmhx9+MH/7299MtWrVzMyZM40xxjzyyCMmPDzcfP311yYtLc107drVNGzY0BQUFBhjjElKSjJOp9NVX1pamnnvvffM9u3bza5du8zo0aONr6+v2bdvn6tMaGio8fPzM3/961/N7t27ze7du83evXuNJJOammrOnDljFi5caCSZjIwMk52dbY4dO2ZOnjxpnE6n+eSTT1x1HTp0yHh7e5uvvvrqksd58uRJM2LECNO0aVOTnZ1tsrOzzcmTJ6/jJ+kZP/30k/Hy8jKTJk0ye/fuNdu3bzfvvvuuOX78uOnYsaNJSEhwK9+jRw/z5JNPuuZDQ0NNzZo1TWJiosnIyDBTpkwxlStXNitWrHAr89ZbbxljjCksLDTt27c3Dz/8sNmyZYvZtWuXGTFihPH39zdHjhwxxhjz+eefm8qVK5tXX33V7Ny506SlpZnXXnvNGGPMkSNHTL169cz48eNdn7MxRX9OIiIizMsvv+zW9oiICDNq1ChjjDGrV682kszRo0dNfn6+mTx5svHz83PVefz4cbN//35TqVIl8+2337rq2LZtm3E4HGbPnj2X/FyPHz9uevXqZbp16+aqMz8///InpAzdLH3VGGMWLFhgvL29zfvvv29++OEHM3r0aFOzZk3TsmXL6/Nhegj91TP91RhjJk6caJxOp/n73/9udu7caQYMGGBq1qxpevTocdltbwQCTjnUsWNHEx4ebgoLC13LXnzxRRMeHm527dplJJn169e71h0+fNhUrVrV9Yvr4o5QnCZNmpipU6e65kNDQ03Pnj3dylz4S9MY9w5zoaFDh5ru3bu75idPnmwaNGjg1v6SjBkzptz/grxYSkqKkWQyMzOLrLvSX5jdunVzKxMbG+v2GV74C3PVqlXGz8/PnDp1ym2b22+/3cyYMcMYY0xUVJTp27dviW2+sL7zLv45mTRpkmnQoIFrPiMjw0gy33//vTGm6Pkv6eese/fuZujQoa75YcOGmU6dOpXYtgs9+eST5eaX45W4mfpqVFSUGTJkiNuyyMjIct9/6a9Hi93+vGvpr8HBwWbChAmu+dOnT5t69eqVmz7MLapyql27dm5vOo+KitLu3bu1c+dOeXl5KTIy0rXO399fjRo1Unp6erF1/frrrxo5cqSaNGmiW265RTVq1NAPP/ygrKwst3Jt27a9qrbGx8drxYoVOnDggCQpKSlJ/fv3t/ZN7S1bttR9992n5s2b6/HHH9f777+vo0ePlqqOqKioIvMlnb+UlBSdOHFC/v7+qlGjhmvau3ev9uzZI0lKS0vTfffdd3UH9P/17t1b+/bt06ZNmyRJH3/8sVq1aqUmTZqUqp74+HjNmzdPp06d0unTp/Xxxx/r97///TW1rTy7Wfpqenp6sT+35R399dKutr/m5uYqOzvb7bPx8vK66p9NT2CQsSWMMSX+knrhhRe0fPlyvfHGG2rYsKGqVq2q//zP/1RBQYFbuerVq1/Vvlu3bq2WLVtqzpw56tq1q7777jt99tlnV1VXRVC5cmUlJydrw4YNWrFihaZOnarRo0dr8+bNqlSpUpF70KdPn76ieks6f4WFhQoODi52PMr5e/KlGXxYkuDgYHXu3Flz585Vu3btNG/ePA0ePLjU9Tz88MPy8fHR4sWL5ePjo/z8fD322GPX3D5b0FdvLPrrpdncX7mCU06dT+UXzt9xxx1q0qSJzpw5o82bN7vWHTlyRLt27VJ4eHixda1bt079+/fXb3/7WzVv3lxBQUHKzMwsdZu8vb0lnRtYdrGBAwcqKSlJs2bNUpcuXRQSEnLFdRZXX3nncDjUvn17jRs3TqmpqfL29tbixYtVu3ZtZWdnu8qdPXtWO3bsKLJ9cee3cePGxe6rTZs2ysnJkZeXlxo2bOg2BQQESJJatGihVatWldjeK/2c+/btqwULFmjjxo3as2ePevfuXeo6vby89OSTTyopKUlJSUnq3bu3qlWrdtl9l6ad5cnN0lfDw8OLPdaKgP56/fur0+lUcHCw22dz5swZpaSkXHbbG4WAU07t379fw4cPV0ZGhubNm6epU6cqISFBd9xxh3r06KH4+Hh988032rZtm5544gnddttt6tGjR7F1NWzYUIsWLVJaWpq2bdumPn36qLCwsNRtCg0NlcPh0Oeff65Dhw7pxIkTrnV9+/bVgQMH9P7775fqdkRYWJj27t2rtLQ0HT58uMg3PMqjzZs36y9/+Yu2bt2qrKwsLVq0SIcOHVJ4eLjuvfdeffHFF/riiy/0ww8/6A9/+EOxzyJZv369Xn/9de3atUvvvvuu/ud//kcJCQnF7q9Lly6KiopSz549tXz5cmVmZmrDhg16+eWXtXXrVknSmDFjNG/ePI0ZM0bp6en67rvv9Prrr7vqCAsL09dff60DBw7o8OHDJR7bo48+qry8PA0dOlSdO3fWbbfdVmLZsLAwnThxQqtWrdLhw4d18uRJ17qBAwfqq6++0pdfflnqn4ft27crIyNDhw8fvuL/TZelm6WvJiQkaNasWZo1a5Z27dqlMWPG6Pvvvy912240+uv/1Xm9+2tCQoImTJigxYsXX/LzKzNlOwQIxenYsaP5wx/+YIYMGWL8/PxMrVq1zEsvveQaCPjLL7+YuLg443Q6TdWqVU3Xrl3Nrl27XNtfPJhs7969pnPnzqZq1aomJCTEvPPOO0UG1xU3qO3igYvGGDN+/HgTFBRkHA6H20A8Y4yJi4szt956a5HBdZdy6tQp89hjj5lbbrnFSDJJSUlXvG1Z2blzp+nataupXbu28fHxMXfeeadrEGhBQYEZOnSoufXWW02dOnVMYmJisYMWx40bZ3r16mWqVatmAgMDzeTJk932cfH5yMvLM3/84x9N3bp1TZUqVUxISIjp27evycrKcpVZuHChadWqlfH29jYBAQHm0Ucfda3buHGjadGihfHx8THnu31Jgw4ff/xxI8nMmjXLbXlxA1eHDBli/P39jSQzZswYt/IdOnQwTZo0uZKP1OXgwYPm/vvvNzVq1DCSzOrVq0u1/Y12M/VVY4x57bXXTEBAgKlRo4Z58sknzciRI8v9IGP661HXsuvdX0+fPm0SEhKMn5+fueWWW8zw4cNNv379ys0gY94mXg516tRJrVq18vijv6+3+++/X+Hh4ZoyZUpZN6XCCw4O1p///GcNHDiwrJtyVYwxaty4sQYPHqzhw4eXdXM8hr4Kif5aXjHIGNfsl19+0YoVK/TVV1/pnXfeKevmVGgnT57U+vXr9fPPP6tp06Zl3ZyrcvDgQX300Uc6cOCAnnrqqbJuDi5AX72+6K/lGwEH16xNmzY6evSoJk6cqEaNGrmta9q0qfbt21fsdjNmzFDfvn1vRBMrjJkzZ+rPf/6zhg0bViG+glucwMBABQQEaObMmapVq5bbuho1apS43ZdffqkOHTp4unk3Nfrq9UV/Ld/9lVtU8Kh9+/aVOFD0/CPfcfP48ccfS1x32223XZevz+Lq0FdxsYreXwk4AADAOnxNHAAAWIeAAwAArEPAAQAA1iHgAAAA6xBwgAps7NixatWqlWu+f//+6tmz5zXVeT3quJl16tRJw4YNc82HhYVVuAcBAjYg4ADXWf/+/eVwOORwOFSlShU1aNBAzz//vH799VeP7/vtt9/W7Nmzr6hsZmamHA6H0tLSrrqO6+nisHbesWPH5HA4in07c2mEhYW5zkvVqlXVuHFj/fWvfy3yNunrbcuWLRo0aJBr3uFw6H//93+v6z7On8vzU82aNdW0aVM9/fTT2r1793XdF1BR8KA/wAO6deumpKQknT59WuvWrdPAgQP166+/avr06UXKnj59WlWqVLku+3U6neWijvJq/Pjxio+P16lTp7Ry5UoNHTpUfn5+Gjx4sMf2Wbt2bY/VfbGVK1eqadOmOnnypL777ju9/fbbatmypT777DPdd999N6wdQHnAFRzAA3x8fBQUFKSQkBD16dNHffv2df2v/fyVilmzZqlBgwby8fGRMUa5ubkaNGiQ6tSpIz8/P917773atm2bW70TJkxwPXRtwIABOnXqlNv6i28vFRYWauLEiWrYsKF8fHz0H//xH3rttdckSfXr15cktW7dWg6HQ506dSq2jvz8fD377LOqU6eOfH199Zvf/EZbtmxxrV+zZo0cDodWrVqltm3bqlq1aoqOjlZGRsZ1+jTdnd/fF198oZYtW8rX11eRkZH67rvvLrttzZo1FRQUpLCwMA0cOFAtWrTQihUrXOsLCgo0cuRI3XbbbapevboiIyPdrhwdOXJEv/vd71SvXj1Vq1ZNzZs317x58y65zwtvUYWFhUmSfvvb38rhcCgsLEyZmZmqVKmS603T502dOlWhoaGlusLk7++voKAgNWjQQD169NDKlSsVGRmpAQMG6OzZs65y06dP1+233y5vb281atRIH330kWvdiBEj9PDDD7vmJ0+e7Pq8z2vUqJFmzJgh6f9+Xt544w0FBwfL399fTz/9dIV4EzzsRsABboCqVau6/cL/8ccf9cknn2jhwoWuW0QPPvigcnJytHTpUqWkpKhNmza677779Msvv0iSPvnkE40ZM0avvfaatm7dquDgYE2bNu2S+x01apQmTpyoV155RTt37tTcuXMVGBgoSfr2228lnftff3Z2thYtWlRsHSNHjtTChQv14Ycf6h//+IcaNmyorl27utp13ujRo/Xmm29q69at8vLy0u9//3vXuvO3UK71NtOFXnjhBb3xxhvasmWL6tSpo0ceeeSK/6gaY7RmzRqlp6e7XT176qmntH79es2fP1/bt2/X448/rm7durlu85w6dUoRERH6/PPPtWPHDg0aNEhxcXHavHnzFe33fDBMSkpSdna2tmzZorCwMHXp0kVJSUluZZOSkly3O8PCwjR27Ngr2seFKlWqpISEBO3bt08pKSmSpMWLFyshIUEjRozQjh07NHjwYD311FNavXq1pHNjiNatW6fCwkJJ0tq1axUQEKC1a9dKknJycrRr1y517NjRtZ/Vq1drz549Wr16tT788EPNnj27TG5zAm7K5iXmgL2efPJJ06NHD9f85s2bjb+/v+nVq5cxxpgxY8aYKlWqmIMHD7rKrFq1yvj5+ZlTp0651XX77bebGTNmGGOMiYqKMkOGDHFbHxkZaVq2bFnsvvPy8oyPj495//33i23n3r17jSSTmppaYvtPnDhhqlSpYj7++GPX+oKCAlO3bl3z+uuvG2OMWb16tZFkVq5c6SrzxRdfGEnm3//+tzHGmH/961+mUaNGZvPmzcW25fzncuGxnHf06FEjyaxevdptf/Pnz3eVOXLkiKlatapZsGBBifWHhoYab29vU716dVOlShUjyfj6+pr169cbY4z58ccfjcPhMAcOHHDb7r777jOjRo0qsd4HHnjAjBgxwjXfsWNHk5CQ4Lbft956yzUvySxevNitjgULFphatWq5zn9aWppxOBxm7969xhhj7r33XjN16tQS21DSuTTGmPT0dCPJ9dlER0eb+Ph4tzKPP/64eeCBB4wxxhw7dsxUqlTJbN261RQWFhp/f3+TmJho7rrrLmOMMXPnzjWBgYGubZ988kkTGhpqzpw541ZfbGxsie0FbgSu4AAe8Pnnn6tGjRry9fVVVFSU7rnnHk2dOtW1PjQ01G1sRkpKik6cOCF/f3/VqFHDNe3du1d79uyRJKWnpxd5od+lXvCXnp6u/Pz8axp7sWfPHp0+fVrt27d3LatSpYruvvtupaenu5Vt0aKF69/BwcGSzr2pWDr33poffvhBd99991W35WIXHvutt96qRo0aFWnTxV544QWlpaVp7dq16ty5s0aPHq3o6GhJ0j/+8Q8ZY3TnnXe6nYO1a9e6zsHZs2f12muvqUWLFq5ztWLFCmVlZV3TsfTs2VNeXl5avHixJGnWrFnq3Lmz65bWqlWr9Mwzz1xV3eb/3+JyOBySzv1cXHg+Jal9+/auz87pdKpVq1Zas2aNvvvuO1WqVEmDBw/Wtm3bdPz4ca1Zs8bt6o107kWdlStXds0HBwe7zj1QVhhkDHhA586dNX36dFWpUkV169YtMoi4evXqbvOFhYUKDg4u9hbOLbfcclVtuB4vwrv4j+OFyy9eduExnl93/jbHlfDz81Nubm6R5ceOHZN0ZYOfL27TxQICAtSwYUM1bNhQCxcuVMOGDdWuXTt16dJFhYWFqly5slJSUtz+WEv/91blN998U2+99ZYmT56s5s2bq3r16ho2bJgKCgqu8CiL5+3trbi4OCUlJenRRx/V3Llzr9tXy88Hl/NjrqTLn89OnTppzZo18vb2VseOHVWrVi01bdpU69ev15o1a9y+Bi+pyM+3w+Eo1bkHPIErOIAHVK9eXQ0bNlRoaOgVfUOqTZs2ysnJkZeXl+sP8PkpICBAkhQeHq5Nmza5bXfx/IXuuOMOVa1aVatWrSp2vbe3tyS5DT69WMOGDeXt7a1vvvnGtez06dPaunWrwsPDL3tcpdG4cWP961//Uk5OjtvyLVu2qFKlSmrYsKHb8guP/ejRo9q1a5caN258xfurVauW/vjHP+r555+XMUatW7fW2bNndfDgwSLnICgoSJK0bt069ejRQ0888YRatmypBg0alPpr2FWqVCn2Mx84cKBWrlypadOm6fTp03r00UdLVW9xCgsLNWXKFNWvX1+tW7eWdO7n6MLzKUkbNmxwO5/nx+F89dVXrsHnHTt21Pz584uMvwHKKwIOUA506dJFUVFR6tmzp5YvX67MzExt2LBBL7/8suvbNQkJCZo1a5ZmzZqlXbt2acyYMfr+++9LrNPX11cvvviiRo4cqTlz5mjPnj3atGmTPvjgA0lSnTp1VLVqVS1btkw///xzsVdPqlevrqFDh+qFF17QsmXLtHPnTsXHx+vkyZMaMGDAFR/fgQMH1LhxY9fA5uLExMQoPDxcvXv31vr167V37159+umnev755zVkyBDVrFnTrfz48eO1atUq7dixQ/3791dAQECpH1D49NNPKyMjQwsXLtSdd96pvn37ql+/flq0aJH27t2rLVu2aOLEiVq6dKmkc4EvOTlZGzZsUHp6ugYPHlwkkF1OWFiYVq1apZycHB09etS1PDw8XO3atdOLL76o3/3ud25X4O677z698847l637yJEjysnJ0T//+U8tWbJEXbp00bfffqsPPvjAdVXqhRde0OzZs/Xee+9p9+7dmjRpkhYtWqTnn3/eVc8999yj48eP67PPPnMFnE6dOulvf/ubateurSZNmpTqmIGyQMABygGHw6GlS5fqnnvu0e9//3vdeeed6t27tzIzM13feoqNjdWrr76qF198UREREdq3b5+GDh16yXpfeeUVjRgxQq+++qrCw8MVGxvrGhvh5eWlKVOmaMaMGapbt6569OhRbB0TJkzQY489pri4OLVp00Y//vijli9frlq1al3x8Z0+fVoZGRk6efJkiWW8vLy0YsUKNWjQQH379lXTpk310ksvaeDAgZo0aVKx7UpISFBERISys7O1ZMkS11WpK1W7dm3FxcVp7NixKiwsVFJSkvr166cRI0aoUaNGeuSRR7R582aFhIRIOvd5tmnTRl27dlWnTp0UFBRU6lD15ptvKjk5WSEhIa6rKucNGDBABQUFbt9Ak86NhTp8+PBl6+7SpYuCg4PVvHlzvfTSSwoPD9f27dvVuXNnV5mePXvq7bff1l//+lc1bdpUM2bMUFJSkivISOduB7Zu3Vq33nqrK8x06NBBhYWFXL1BheEwxsOP8QSA62jNmjXq3Lmzjh49etXjk8qr1157TfPnz7+iZ/oAuDSu4ABAGTtx4oS2bNmiqVOn6tlnny3r5gBWIOAAQBl75pln9Jvf/EYdO3YscnsKwNXhFhUAALAOV3AAAIB1CDgAAMA6BBwAAGAdAg4AALAOAQcAAFiHgAMAAKxDwAEAANYh4AAAAOsQcAAAgHX+H2z4L+LGru5LAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_names = ['Down', 'Up']\n",
    "def plot_image(i, predictions_array, true_label, news):\n",
    "  true_label, news = true_label[i], news.iloc[i]\n",
    "  plt.grid(False)\n",
    "\n",
    "  bars = news.to_list()[2:6]\n",
    "  plt.bar(['polarity_t', 'subjectivity_t',      \n",
    "           'polarity_d', 'subjectivity_d'], \n",
    "            bars,\n",
    "            )\n",
    "\n",
    "  predicted_label =predictions_array[0]\n",
    "\n",
    "  plt.xlabel(\"Prediction: {} Reality: {}\".format(\n",
    "    class_names[int(predicted_label)],\n",
    "    class_names[target[i]]))\n",
    "\n",
    "plot_image(0, predictions[0], class_names, test_set)"
   ]
  }
 ],
 "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": "3.9.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}