sns_nn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "# SNS Sentiment Analysis NN\n",
    "\n",
    "Analyze social media texts and measure potential inflammatory / offensive language with a feedforward neural network."
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "5614af81228c83e9"
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "import ast\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "import data_util as du\n",
    "import keras_tuner as kt\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout\n",
    "from tensorflow.keras.optimizers.legacy import Adam\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from matplotlib import pyplot as plt"
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "c9f2c98f368e4b25"
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Load Data\n",
    "\n",
    "Load preprocessed data, vocabulary included. "
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "b8b14a0fda3ba9b7"
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8000,) (8000,)\n",
      "(2000,) (2000,)\n",
      "Vocabulary size: 11560\n"
     ]
    }
   ],
   "source": [
    "# Load dataset\n",
    "train = pd.read_csv(\"./datasets/processed/train.csv\")\n",
    "X_train = train[train.columns[0]].apply(ast.literal_eval)\n",
    "y_train = train[train.columns[1]].astype(int)\n",
    "print(X_train.shape, y_train.shape)\n",
    "\n",
    "test = pd.read_csv(\"./datasets/processed/test.csv\")\n",
    "X_test = test[test.columns[0]].apply(ast.literal_eval)\n",
    "y_test = test[test.columns[1]].astype(int)\n",
    "print(X_test.shape, y_test.shape)\n",
    "\n",
    "# Load vocabulary\n",
    "vocabulary = ast.literal_eval(open(\"./datasets/processed/vocabulary.txt\", \"r\").read())\n",
    "print(\"Vocabulary size:\", len(vocabulary))"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-12T23:53:09.912839Z",
     "start_time": "2024-04-12T23:53:09.379755Z"
    }
   },
   "id": "f3353cbd63cea128",
   "execution_count": 2
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Lengths of train and test: 8000 2000\n",
      "train:test ratio: 0.8\n",
      "test:train ratio: 0.2\n"
     ]
    }
   ],
   "source": [
    "train_text = [\" \".join(s) for s in X_train]\n",
    "test_text = [\" \".join(s) for s in X_test]\n",
    "print(\"Lengths of train and test:\", len(train_text), len(test_text))\n",
    "print(\"train:test ratio:\", len(train_text) / (len(train_text) + len(test_text)))\n",
    "print(\"test:train ratio:\", len(test_text) / (len(train_text) + len(test_text)))"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-12T23:53:09.925444Z",
     "start_time": "2024-04-12T23:53:09.913639Z"
    }
   },
   "id": "d678733692490132",
   "execution_count": 3
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Vectorize data\n",
    "\n",
    "In multinomial vectorization, we use it to calculate probabilities based on frequency of words from the vocabulary than whether the word has appeared at least once."
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "e7f386939c8df7b9"
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "# Multinomial vectorizer, used to calculate probabilities based on vocabulary frequencies \n",
    "m_vectorizer = CountVectorizer(input='content', stop_words='english', binary=False, vocabulary=vocabulary, tokenizer=None, preprocessor=None)\n",
    "X_train_m = m_vectorizer.fit_transform(train_text)\n",
    "X_test_m = m_vectorizer.transform(test_text)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-12T23:53:10.054932Z",
     "start_time": "2024-04-12T23:53:09.926155Z"
    }
   },
   "id": "4ca5c2fda149009b",
   "execution_count": 4
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Create a feedforward neural network then train it\n",
    "\n",
    "Network has been tuned with different values for optimal results."
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "78a043c02d2fd1e9"
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "def create_ff_neural_network(X_training_data):\n",
    "    model = Sequential([\n",
    "        # Input layer\n",
    "        # Input shape is also just vocabulary size\n",
    "        Dense(units=256, input_dim=X_training_data.shape[1], activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        Dense(units=128, activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        Dense(units=64, activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        Dense(units=32, activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        # Output layer\n",
    "        Dense(units=1, activation='sigmoid')\n",
    "    ])\n",
    "    \n",
    "    # Legacy Adam\n",
    "    opt = Adam(learning_rate=2e-5)\n",
    "    model.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])\n",
    "    return model"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-12T23:53:14.309037Z",
     "start_time": "2024-04-12T23:53:14.303724Z"
    }
   },
   "id": "699bbea8dd4293a4",
   "execution_count": 5
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense (Dense)               (None, 256)               2959616   \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 256)               0         \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 128)               32896     \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 128)               0         \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 64)                8256      \n",
      "                                                                 \n",
      " dropout_2 (Dropout)         (None, 64)                0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 32)                2080      \n",
      "                                                                 \n",
      " dropout_3 (Dropout)         (None, 32)                0         \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 1)                 33        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 3002881 (11.46 MB)\n",
      "Trainable params: 3002881 (11.46 MB)\n",
      "Non-trainable params: 0 (0.00 Byte)\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "multinomial_model_ff = create_ff_neural_network(X_train_m)\n",
    "multinomial_model_ff.summary()"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-12T23:53:15.064554Z",
     "start_time": "2024-04-12T23:53:14.816930Z"
    }
   },
   "id": "cb36400665ebf981",
   "execution_count": 6
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Learning Rate Scheduler\n",
    "\n",
    "If we have a high learning rate, we may unintentionally speed up convergence by making large updates. By reducing the learning rate, it helps the model fine-tune its weights allowing it to obtain a more accurate stance. This is made to prevent over-fitting with our large data."
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "e21dc7a1ebb5aaf1"
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "# Learning rate scheduler\n",
    "def scheduler(epoch, lr):\n",
    "    if epoch < 5:\n",
    "        return lr\n",
    "    else:\n",
    "        return lr * tf.math.exp(-0.1)\n",
    "\n",
    "# Create a learning rate callback\n",
    "lr_callback = tf.keras.callbacks.LearningRateScheduler(scheduler)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-12T23:53:19.386083Z",
     "start_time": "2024-04-12T23:53:19.382640Z"
    }
   },
   "id": "1994fdbffa2d705b",
   "execution_count": 7
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Multinomial Feedforward Neural Network:\n",
      "Epoch 1/10\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-04-08 20:08:55.925207: W tensorflow/tsl/platform/profile_utils/cpu_utils.cc:128] Failed to get CPU frequency: 0 Hz\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "250/250 [==============================] - 2s 5ms/step - loss: 0.6756 - accuracy: 0.5789 - val_loss: 0.6532 - val_accuracy: 0.6745 - lr: 2.0000e-05\n",
      "Epoch 2/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.6373 - accuracy: 0.6806 - val_loss: 0.6161 - val_accuracy: 0.7100 - lr: 2.0000e-05\n",
      "Epoch 3/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.5923 - accuracy: 0.7253 - val_loss: 0.5780 - val_accuracy: 0.7325 - lr: 2.0000e-05\n",
      "Epoch 4/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.5438 - accuracy: 0.7514 - val_loss: 0.5482 - val_accuracy: 0.7350 - lr: 2.0000e-05\n",
      "Epoch 5/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.5016 - accuracy: 0.7739 - val_loss: 0.5313 - val_accuracy: 0.7390 - lr: 2.0000e-05\n",
      "Epoch 6/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.4661 - accuracy: 0.7929 - val_loss: 0.5245 - val_accuracy: 0.7405 - lr: 1.8097e-05\n",
      "Epoch 7/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.4422 - accuracy: 0.8026 - val_loss: 0.5247 - val_accuracy: 0.7370 - lr: 1.6375e-05\n",
      "Epoch 8/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.4203 - accuracy: 0.8135 - val_loss: 0.5282 - val_accuracy: 0.7405 - lr: 1.4816e-05\n",
      "Epoch 9/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.4064 - accuracy: 0.8204 - val_loss: 0.5325 - val_accuracy: 0.7405 - lr: 1.3406e-05\n",
      "Epoch 10/10\n",
      "250/250 [==============================] - 1s 4ms/step - loss: 0.3930 - accuracy: 0.8296 - val_loss: 0.5388 - val_accuracy: 0.7385 - lr: 1.2131e-05\n"
     ]
    }
   ],
   "source": [
    "print(\"Multinomial Feedforward Neural Network:\")\n",
    "\n",
    "ff_history = multinomial_model_ff.fit(np.array(X_train_m.toarray()),\n",
    "              np.array(y_train),\n",
    "              epochs=10, batch_size=32,\n",
    "              validation_data=(np.array(X_test_m.toarray()), np.array(y_test)),\n",
    "              callbacks=[lr_callback])"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-09T00:09:07.077834Z",
     "start_time": "2024-04-09T00:08:55.160290Z"
    }
   },
   "id": "2cd7fb6d68c6ce84",
   "execution_count": 9
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "# multinomial_model_ff.save('./models/sns_multinomial_ff.h5')\n",
    "# multinomial_model_ff.save('./models/sns_multinomial_ff.keras')"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-09T00:09:09.533429Z",
     "start_time": "2024-04-09T00:09:09.452704Z"
    }
   },
   "id": "33fc96563ddb9a73",
   "execution_count": 10
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "# multinomial_model_ff = keras.models.load_model('./models/sns_multinomial.h5')"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-09T00:09:10.552600Z",
     "start_time": "2024-04-09T00:09:10.447650Z"
    }
   },
   "id": "e3b9c6ea201603d8",
   "execution_count": 11
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Tuned Neural Network"
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "5fa721c1a41d21ca"
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "def create_tuned_ffnn(hp: kt.HyperParameters):\n",
    "    model = Sequential([\n",
    "        # Input layer\n",
    "        # Input shape is also just vocabulary size\n",
    "        Dense(units=X_train_m.shape[1], input_dim=X_train_m.shape[1], activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        Dense(units=128, activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        Dense(units=64, activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        Dense(units=32, activation='tanh'),\n",
    "        Dropout(0.2),\n",
    "        # Output layer\n",
    "        Dense(units=1, activation='sigmoid')\n",
    "    ])\n",
    "    \n",
    "    hp_learning_rate = hp.Choice('learning_rate', values=[1e-2, 1e-3, 1e-4, 1e-5])\n",
    "    # Legacy Adam\n",
    "    opt = Adam(learning_rate=hp_learning_rate)\n",
    "    model.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])\n",
    "    return model"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:21:22.888918Z",
     "start_time": "2024-04-13T00:21:22.877992Z"
    }
   },
   "id": "db1d96bfb6b95948",
   "execution_count": 30
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/tuner0.json\n"
     ]
    }
   ],
   "source": [
    "tuner = kt.RandomSearch(create_tuned_ffnn,\n",
    "                     objective='val_accuracy',\n",
    "                     max_trials=5,\n",
    "                     executions_per_trial=1,\n",
    "                     project_name='tuners',\n",
    "                     directory='./')"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:21:35.563433Z",
     "start_time": "2024-04-13T00:21:35.549136Z"
    }
   },
   "id": "30c3bee73cedab26",
   "execution_count": 33
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "stop_early = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:21:36.398709Z",
     "start_time": "2024-04-13T00:21:36.396625Z"
    }
   },
   "id": "379bf96234ce93be",
   "execution_count": 34
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trial 15 Complete [00h 05m 19s]\n",
      "val_accuracy: 0.7444999814033508\n",
      "\n",
      "Best val_accuracy So Far: 0.7444999814033508\n",
      "Total elapsed time: 00h 32m 29s\n"
     ]
    }
   ],
   "source": [
    "tuner.search(np.array(X_train_m.toarray()), np.array(y_train), validation_data=(np.array(X_test_m.toarray()), np.array(y_test)), epochs=5, callbacks=[stop_early])"
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "10dab9092677555",
   "execution_count": 36
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "best_params = tuner.get_best_hyperparameters(num_trials=1)[0]"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:27:02.029974Z",
     "start_time": "2024-04-13T00:27:02.026964Z"
    }
   },
   "id": "eba398276088dd89",
   "execution_count": 37
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "best_nn = tuner.hypermodel.build(best_params)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:27:04.870129Z",
     "start_time": "2024-04-13T00:27:04.343083Z"
    }
   },
   "id": "c8f455656d349244",
   "execution_count": 38
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "250/250 [==============================] - 44s 172ms/step - loss: 0.6397 - accuracy: 0.6639 - val_loss: 0.5932 - val_accuracy: 0.7330 - lr: 1.0000e-05\n",
      "Epoch 2/10\n",
      "250/250 [==============================] - 43s 169ms/step - loss: 0.5600 - accuracy: 0.7464 - val_loss: 0.5480 - val_accuracy: 0.7370 - lr: 1.0000e-05\n",
      "Epoch 3/10\n",
      "250/250 [==============================] - 41s 163ms/step - loss: 0.5032 - accuracy: 0.7781 - val_loss: 0.5252 - val_accuracy: 0.7410 - lr: 1.0000e-05\n",
      "Epoch 4/10\n",
      "250/250 [==============================] - 41s 163ms/step - loss: 0.4621 - accuracy: 0.7960 - val_loss: 0.5174 - val_accuracy: 0.7425 - lr: 1.0000e-05\n",
      "Epoch 5/10\n",
      "250/250 [==============================] - 40s 159ms/step - loss: 0.4257 - accuracy: 0.8179 - val_loss: 0.5205 - val_accuracy: 0.7415 - lr: 1.0000e-05\n",
      "Epoch 6/10\n",
      "250/250 [==============================] - 39s 157ms/step - loss: 0.3971 - accuracy: 0.8313 - val_loss: 0.5291 - val_accuracy: 0.7390 - lr: 9.0484e-06\n",
      "Epoch 7/10\n",
      "250/250 [==============================] - 39s 156ms/step - loss: 0.3769 - accuracy: 0.8405 - val_loss: 0.5407 - val_accuracy: 0.7390 - lr: 8.1873e-06\n",
      "Epoch 8/10\n",
      "250/250 [==============================] - 39s 155ms/step - loss: 0.3584 - accuracy: 0.8479 - val_loss: 0.5539 - val_accuracy: 0.7365 - lr: 7.4082e-06\n",
      "Epoch 9/10\n",
      "250/250 [==============================] - 42s 170ms/step - loss: 0.3433 - accuracy: 0.8543 - val_loss: 0.5670 - val_accuracy: 0.7350 - lr: 6.7032e-06\n",
      "Epoch 10/10\n",
      "250/250 [==============================] - 42s 168ms/step - loss: 0.3314 - accuracy: 0.8545 - val_loss: 0.5800 - val_accuracy: 0.7365 - lr: 6.0653e-06\n"
     ]
    }
   ],
   "source": [
    "best_nn_history = best_nn.fit(np.array(X_train_m.toarray()),\n",
    "              np.array(y_train),\n",
    "              epochs=10, batch_size=32,\n",
    "              validation_data=(np.array(X_test_m.toarray()), np.array(y_test)),\n",
    "              callbacks=[lr_callback])"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:33:59.167992Z",
     "start_time": "2024-04-13T00:27:07.694131Z"
    }
   },
   "id": "6d0e7c6cbed9d6b",
   "execution_count": 39
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "best_nn.save('./models/sns_tuned_nn.h5')\n",
    "best_nn.save_weights('./models/sns_tuned_nn.weights.h5')"
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "dd36f15d40fcfcca",
   "execution_count": null
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Make predictions"
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "a0ebecda2c08f996"
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "def probs_to_preds(probabilities: list) -> list[int]:\n",
    "    \"\"\"\n",
    "    Converts continuous (sigmoid) outputs to discrete binary probabilities\n",
    "    :param probabilities: List of probabilities between [0, 1]\n",
    "    :return: \n",
    "    \"\"\"\n",
    "    return [1 if p[0] > 0.5 else 0 for p in probabilities]"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:35:12.378439Z",
     "start_time": "2024-04-13T00:35:12.372542Z"
    }
   },
   "id": "a657ffdf2cc5425d",
   "execution_count": 40
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 0s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "raw_predictions_ff = multinomial_model_ff.predict(X_test_m.toarray())\n",
    "preds_ff = probs_to_preds(raw_predictions_ff)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-09T00:10:04.292739Z",
     "start_time": "2024-04-09T00:10:04.064873Z"
    }
   },
   "id": "1978eb826196376e",
   "execution_count": 15
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 1s 14ms/step\n"
     ]
    }
   ],
   "source": [
    "raw_predictions_best_nn = best_nn.predict(X_test_m.toarray())\n",
    "preds_best_nn = probs_to_preds(raw_predictions_best_nn)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:35:18.236816Z",
     "start_time": "2024-04-13T00:35:17.071222Z"
    }
   },
   "id": "40e61d0befa97b30",
   "execution_count": 41
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision: 0.718139534883721\n",
      "Recall: 0.7774420946626385\n",
      "F1 Score: 0.746615087040619\n",
      "Accuracy: 0.738\n"
     ]
    }
   ],
   "source": [
    "_, _, _, _ = du.get_prfa(y_test, preds_ff, verbose=True)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-09T00:10:11.081031Z",
     "start_time": "2024-04-09T00:10:11.045935Z"
    }
   },
   "id": "85684e6910a7c365",
   "execution_count": 16
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision: 0.71939736346516\n",
      "Recall: 0.7693856998992951\n",
      "F1 Score: 0.743552311435523\n",
      "Accuracy: 0.7365\n"
     ]
    }
   ],
   "source": [
    "_ = du.get_prfa(y_test, preds_best_nn, verbose=True)"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:35:27.893422Z",
     "start_time": "2024-04-13T00:35:27.881959Z"
    }
   },
   "id": "65a12529a348b04c",
   "execution_count": 43
  },
  {
   "cell_type": "code",
   "outputs": [],
   "source": [
    "def predict_inflammatory_level(model: Sequential, text):\n",
    "    # Vectorize the input text\n",
    "    X_input = m_vectorizer.transform([text])\n",
    "    \n",
    "    # Get the prediction from the trained model\n",
    "    prediction = model.predict(X_input.toarray())\n",
    "    \n",
    "    # The output is a probability value between 0 and 1\n",
    "    inflammatory_level = prediction[0][0]\n",
    "    \n",
    "    return inflammatory_level"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:35:41.103160Z",
     "start_time": "2024-04-13T00:35:41.098370Z"
    }
   },
   "id": "d31df7dc6b825378",
   "execution_count": 45
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 13ms/step\n",
      "0.23211204\n"
     ]
    }
   ],
   "source": [
    "# Example of predicting inflammatory text with a probability\n",
    "print(predict_inflammatory_level(multinomial_model_ff, \"test text\"))"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-09T00:10:59.928415Z",
     "start_time": "2024-04-09T00:10:59.889477Z"
    }
   },
   "id": "c0d00bd087f65adf",
   "execution_count": 18
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 12ms/step\n",
      "0.19137672\n"
     ]
    }
   ],
   "source": [
    "print(predict_inflammatory_level(best_nn, \"test text, very passive\"))"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:36:10.418261Z",
     "start_time": "2024-04-13T00:36:10.357500Z"
    }
   },
   "id": "1c549101fa7e5632",
   "execution_count": 49
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Plot history"
   ],
   "metadata": {
    "collapsed": false
   },
   "id": "cd0c240c167bcc95"
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 500x800 with 0 Axes>"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5, 8))\n",
    "pd.DataFrame(ff_history.history).plot()\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.savefig('./graphs/ff_history_2.png')\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-07T17:43:31.867423Z",
     "start_time": "2024-04-07T17:43:31.704734Z"
    }
   },
   "id": "fe570515a32b7956",
   "execution_count": 271
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 0 Axes>"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 500x800 with 0 Axes>"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 640x480 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.clf()\n",
    "plt.figure(figsize=(5, 8))\n",
    "pd.DataFrame(best_nn_history.history).plot()\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.title(\"Tuned Neural Network Performance over Epochs\")\n",
    "plt.savefig('./graphs/best_nn_history.png')\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-04-13T00:39:49.004252Z",
     "start_time": "2024-04-13T00:39:48.804293Z"
    }
   },
   "id": "d937544e7b5e53d0",
   "execution_count": 53
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}