BERTModel_Final.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "dd74d1ac-92c4-41bb-9879-3352eeef6a82",
   "metadata": {},
   "source": [
    "## BERT for Pharmacy Website Text Classification\n",
    "### Goal: Train model to determine if online pharmacy is accredited or not.\n",
    "#### Dataset: \n",
    "- Accredited Online Pharmacies are labeled 0\n",
    "- Non-Accredited Online Pharmacies are labeled 1 \n",
    "- 'Text' column created in this notebook is a combination of body, footer, and header text scraped from each website\n",
    "- This is a subset of websites. The final dataset will be larger."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3319c0da-f02d-4d1c-b1c6-36a4b767b102",
   "metadata": {},
   "source": [
    "### Text Pre-Processing for Feature-Based Training\n",
    "- BERT (Bidirectional Encoder Representations from Transformers)\n",
    "- BERT makes use of a Transformer, an attention mechanism that learns contextual relations between words (or sub-words) in a text\n",
    "- Transformer includes two separate mechanisms — an encoder that reads the text input and a decoder that produces a prediction for the task. \n",
    "- Since BERT’s goal is to generate a language model, only the encoder mechanism is necessary\n",
    "- the Transformer encoder reads the entire sequence of words at once\n",
    "\n",
    "- Resources: \n",
    "    - https://keras.io/examples/nlp/text_extraction_with_bert/#preprocess-the-data\n",
    "    - https://collab.its.virginia.edu/access/lessonbuilder/item/2079372/group/25e8dc9b-3e66-4249-ae38-3a124dead1e4/Module%207:%20Seque_%2010_12%20-%2010_18/7.5%20%20Video%20+%20Qu_%20and%20Attention/M7.5_%20NLP%20with%20Attention,%20Transformers,%20and%20BERT.pdf\n",
    "    - https://towardsdatascience.com/multi-label-text-classification-using-bert-and-tensorflow-d2e88d8f488d#98ee\n",
    "    - https://curiousily.com/posts/multi-label-text-classification-with-bert-and-pytorch-lightning/\n",
    "    - https://www.tensorflow.org/text/tutorials/classify_text_with_bert\n",
    "    - https://swatimeena989.medium.com/bert-text-classification-using-keras-903671e0207d\n",
    "    - https://towardsdatascience.com/how-to-train-a-bert-model-from-scratch-72cfce554fc6\n",
    "    - https://huggingface.co/docs/tokenizers/pipeline"
   ]
  },
  {
   "attachments": {
    "371ae26e-6063-4cb6-b68b-0ea720530d26.jpg": {
     "image/jpeg": "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"
    }
   },
   "cell_type": "markdown",
   "id": "f5b0e530-6716-4573-a6c3-4d3a076e17d5",
   "metadata": {},
   "source": [
    "![bertmodel.jpg](attachment:371ae26e-6063-4cb6-b68b-0ea720530d26.jpg)\n",
    "\n",
    "Two-Phase training:\n",
    "\n",
    "1.) Pretrain to understand language\n",
    "\n",
    "2.) Finetune to learn a specific task\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6fe9c3a0-9c58-4437-a06f-d000812c9105",
   "metadata": {},
   "source": [
    "### Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "5f0db4fc-a724-4f48-99c4-22a768cbc44a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#import libraries\n",
    "import tensorflow as tf\n",
    "import tensorflow_hub as hub\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "import numpy as np\n",
    "import re\n",
    "import unicodedata\n",
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "import keras\n",
    "from tqdm import tqdm\n",
    "import pickle\n",
    "from keras.models import Model\n",
    "import keras.backend as K\n",
    "from sklearn.metrics import confusion_matrix,f1_score,classification_report\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "import itertools\n",
    "from keras.models import load_model\n",
    "from sklearn.utils import shuffle\n",
    "from transformers import *\n",
    "from transformers import BertTokenizer, TFBertModel, BertConfig\n",
    "import tensorflow_text as text\n",
    "\n",
    "from tensorflow.keras.layers import Input, Conv2D, Dense, Flatten, Dropout\n",
    "from tensorflow.keras.layers import GlobalMaxPooling2D, MaxPooling2D\n",
    "from tensorflow.keras.layers import BatchNormalization\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.models import load_model\n",
    "\n",
    "\n",
    "# to make this notebook's output stable across runs\n",
    "np.random.seed(42)\n",
    "tf.random.set_seed(42)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec4eb449-0ac8-4814-a747-4db40348d7c2",
   "metadata": {},
   "source": [
    "### Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "7a2699bd-3cf7-4a59-b1b5-b3bb400276df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "72"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#import data\n",
    "pharm_data = pd.read_csv('pharmacy_dataset_reduced.csv')\n",
    "data = pharm_data.copy()\n",
    "data.head(2)\n",
    "len(data)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4dbfabd5-e25e-4923-8675-6ed097dca9ee",
   "metadata": {},
   "source": [
    "### Preprocess & Clean Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "092c4dd8-3f67-4844-8577-c385b4d0b63a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def unicode_to_ascii(s):\n",
    "    return ''.join(c for c in unicodedata.normalize('NFD', s) if unicodedata.category(c) != 'Mn')\n",
    "\n",
    "def clean_stopwords_shortwords(w):\n",
    "    stopwords_list=stopwords.words('english')\n",
    "    words = w.split() \n",
    "    clean_words = [word for word in words if (word not in stopwords_list) and len(word) > 2]\n",
    "    return \" \".join(clean_words) \n",
    "\n",
    "def preprocess_sentence(w):\n",
    "    w = unicode_to_ascii(w.lower().strip())\n",
    "    w = re.sub(r\"([?.!,¿])\", r\" \", w)\n",
    "    w = re.sub(r'[\" \"]+', \" \", w)\n",
    "    w = re.sub(r\"[^a-zA-Z?.!,¿]+\", \" \", w)\n",
    "    w=clean_stopwords_shortwords(w)\n",
    "    w=re.sub(r'@\\w+', '',w)\n",
    "    return w"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "36e06a85-2762-4b97-8e15-e30f7892fccc",
   "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>Website</th>\n",
       "      <th>Body</th>\n",
       "      <th>Header</th>\n",
       "      <th>Footer</th>\n",
       "      <th>Image Urls</th>\n",
       "      <th>Accredited</th>\n",
       "      <th>Zipcode</th>\n",
       "      <th>Valid_phone</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>https://acariahealth.envolvehealth.com</td>\n",
       "      <td>Skip to Main Content Home Contact Insights Sea...</td>\n",
       "      <td>Home Contact Insights Search Search About Over...</td>\n",
       "      <td>Locations Referral Forms Careers Disclaimer HI...</td>\n",
       "      <td>['https://acariahealth.envolvehealth.com/conte...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Skip to Main Content Home Contact Insights Sea...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>https://alto.com</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "      <td>Our Story Careers Drive for Alto For Providers...</td>\n",
       "      <td>['https://images.prismic.io/alto/176341aa-258d...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  Website  \\\n",
       "0  https://acariahealth.envolvehealth.com   \n",
       "1                        https://alto.com   \n",
       "\n",
       "                                                Body  \\\n",
       "0  Skip to Main Content Home Contact Insights Sea...   \n",
       "1  Shop Alto Essentials to add pharmacy products ...   \n",
       "\n",
       "                                              Header  \\\n",
       "0  Home Contact Insights Search Search About Over...   \n",
       "1  Shop Alto Essentials to add pharmacy products ...   \n",
       "\n",
       "                                              Footer  \\\n",
       "0  Locations Referral Forms Careers Disclaimer HI...   \n",
       "1  Our Story Careers Drive for Alto For Providers...   \n",
       "\n",
       "                                          Image Urls  Accredited  Zipcode  \\\n",
       "0  ['https://acariahealth.envolvehealth.com/conte...           0        0   \n",
       "1  ['https://images.prismic.io/alto/176341aa-258d...           0        0   \n",
       "\n",
       "   Valid_phone                                               text  \n",
       "0            1  Skip to Main Content Home Contact Insights Sea...  \n",
       "1            1  Shop Alto Essentials to add pharmacy products ...  "
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#combine columns into new column containing all text\n",
    "data['text'] = data['Body'].astype(str) + data['Header'].astype(str) + data['Footer'].astype(str)\n",
    "data.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "23e44509-6427-445d-b93d-9c4238c5cd88",
   "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>Website</th>\n",
       "      <th>Body</th>\n",
       "      <th>Header</th>\n",
       "      <th>Footer</th>\n",
       "      <th>Image Urls</th>\n",
       "      <th>label</th>\n",
       "      <th>Zipcode</th>\n",
       "      <th>Valid_phone</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>https://acariahealth.envolvehealth.com</td>\n",
       "      <td>Skip to Main Content Home Contact Insights Sea...</td>\n",
       "      <td>Home Contact Insights Search Search About Over...</td>\n",
       "      <td>Locations Referral Forms Careers Disclaimer HI...</td>\n",
       "      <td>['https://acariahealth.envolvehealth.com/conte...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Skip to Main Content Home Contact Insights Sea...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>https://alto.com</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "      <td>Our Story Careers Drive for Alto For Providers...</td>\n",
       "      <td>['https://images.prismic.io/alto/176341aa-258d...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  Website  \\\n",
       "0  https://acariahealth.envolvehealth.com   \n",
       "1                        https://alto.com   \n",
       "\n",
       "                                                Body  \\\n",
       "0  Skip to Main Content Home Contact Insights Sea...   \n",
       "1  Shop Alto Essentials to add pharmacy products ...   \n",
       "\n",
       "                                              Header  \\\n",
       "0  Home Contact Insights Search Search About Over...   \n",
       "1  Shop Alto Essentials to add pharmacy products ...   \n",
       "\n",
       "                                              Footer  \\\n",
       "0  Locations Referral Forms Careers Disclaimer HI...   \n",
       "1  Our Story Careers Drive for Alto For Providers...   \n",
       "\n",
       "                                          Image Urls  label  Zipcode  \\\n",
       "0  ['https://acariahealth.envolvehealth.com/conte...      0        0   \n",
       "1  ['https://images.prismic.io/alto/176341aa-258d...      0        0   \n",
       "\n",
       "   Valid_phone                                               text  \n",
       "0            1  Skip to Main Content Home Contact Insights Sea...  \n",
       "1            1  Shop Alto Essentials to add pharmacy products ...  "
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#rename Accredited as 'label'\n",
    "data=data.rename(columns = {'Accredited': 'label'}, inplace = False)\n",
    "data.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "1c71ee3f-172d-4b30-9850-ce6377b60a02",
   "metadata": {},
   "outputs": [],
   "source": [
    "#reduce dataset to only the columns we need, 'label' and 'text'\n",
    "data = data.drop(['Website', 'Body', 'Header', 'Footer', 'Image Urls', 'Zipcode', 'Valid_phone'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "01cb61cb-b3be-4b24-8c62-0f43647dda4c",
   "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>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Skip to Main Content Home Contact Insights Sea...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Shop Alto Essentials to add pharmacy products ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Loading... Skip to navigation Skip to main con...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label                                               text\n",
       "0      0  Skip to Main Content Home Contact Insights Sea...\n",
       "1      0  Shop Alto Essentials to add pharmacy products ...\n",
       "2      0  Loading... Skip to navigation Skip to main con..."
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Sentences contain the entire text data and labels contain all the corresponding labels\n",
    "data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "3fbfcd1d-4616-4d13-8987-d8160ec37f6d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#drop any missing values\n",
    "data = data.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "270be1ce-3456-4e70-92f0-fb76d8acd24c",
   "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>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Loading... × Close alert Fly with Birdi!  Free...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>1</td>\n",
       "      <td>Log In Register Checkout Languages: Currencies...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>Looking to save up to 80% on your medications?...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    label                                               text\n",
       "4       0  Loading... × Close alert Fly with Birdi!  Free...\n",
       "62      1  Log In Register Checkout Languages: Currencies...\n",
       "18      0  Looking to save up to 80% on your medications?..."
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#reset index\n",
    "data=data.reset_index(drop=True) \n",
    "#shuffle data\n",
    "data = shuffle(data)   \n",
    "data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "6272327a-f1ff-456a-ac15-b0668ae1e234",
   "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>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>loading close alert fly birdi free shipping or...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>1</td>\n",
       "      <td>log register checkout languages currencies dol...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>looking save medications find mygeniusrx healt...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    label                                               text\n",
       "4       0  loading close alert fly birdi free shipping or...\n",
       "62      1  log register checkout languages currencies dol...\n",
       "18      0  looking save medications find mygeniusrx healt..."
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#apply the function to clean the text - all lower case, strip spaces and characters \n",
    "data['text']=data['text'].map(preprocess_sentence)    \n",
    "data.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f23c128-d30e-4626-bcf4-b47c87da188a",
   "metadata": {},
   "source": [
    "### Split Data - Training & Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "42cc4290-89d8-4f5c-971a-2fa6028827a6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#define the number of unique classes under 'label'\n",
    "num_classes=len(data.label.unique())\n",
    "num_classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "0c353831-5872-4ca6-8039-239f72b59e2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#define y variable\n",
    "y = tf.keras.utils.to_categorical(data[\"label\"].values, num_classes=num_classes)\n",
    "#y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "ae372ae7-f20d-4ca9-b488-7d8096c309e5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(50, 22, 50, 22)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#split data into test and train\n",
    "x_train, x_test, y_train, y_test = train_test_split(data['text'], y, test_size=0.3)\n",
    "len(x_train), len(x_test), len(y_train), len(y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aedeb195-7148-49fd-91bb-0ad28be047ec",
   "metadata": {},
   "source": [
    "### Data Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da918d68-197a-46d5-8762-4f223be4f986",
   "metadata": {},
   "source": [
    "#### Load BERT with TensorFlow Hub\n",
    "- repository of trained machine learning models\n",
    "- use universal-sentence-encoder-cmlm/multilingual-base = universal sentence encoder (100+ languages)\n",
    "    - uses conditional masked language model\n",
    "- Goal: turn text into high-dim vectors that capture sentence-level semantics\n",
    "    - get embeddings from input text with preprocessor and encoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "b83d7471-d20c-42e0-a067-2ffba9b55953",
   "metadata": {},
   "outputs": [],
   "source": [
    "#load bert with tensorflow hub\n",
    "preprocessor = hub.KerasLayer(\"https://tfhub.dev/google/universal-sentence-encoder-cmlm/multilingual-preprocess/2\")\n",
    "encoder = hub.KerasLayer(\"https://tfhub.dev/google/universal-sentence-encoder-cmlm/multilingual-base/1\")\n",
    "\n",
    "#function to get word embeddings\n",
    "def get_embeddings(sentences):\n",
    "  '''return BERT-like embeddings of input text\n",
    "  Args:\n",
    "    - sentences: list of strings\n",
    "  Output:\n",
    "    - BERT-like embeddings: tf.Tensor of shape=(len(sentences), 768)\n",
    "  '''\n",
    "  preprocessed_text = preprocessor(sentences)\n",
    "  return encoder(preprocessed_text)['pooled_output']\n",
    "\n",
    "#test function\n",
    "#get_embeddings([\n",
    "   # \"This is a test para ver el futuro of the model.\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "442c59b1-edd3-4dcc-b4e6-4a6c9a48ed94",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Note: an option in the future is to observe semantic textual similarities"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ecabf30-0907-479b-a159-3d667d8e727a",
   "metadata": {},
   "source": [
    "### Create & Train Classification Model\n",
    "- Observe different metrics during training: Precision, Recall, F1 Score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "4d23d5ab-e025-4a1d-b69b-420e88b273aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "#resource: https://towardsdatascience.com/multi-label-text-classification-using-bert-and-tensorflow-d2e88d8f488d#98ee\n",
    "from keras import backend as K\n",
    "\n",
    "#functions to find recall\n",
    "def balanced_recall(y_true, y_pred):\n",
    "    \"\"\"This function calculates the balanced recall metric\n",
    "    recall = TP / (TP + FN)\n",
    "    \"\"\"\n",
    "    recall_by_class = 0\n",
    "    # iterate over each predicted class to get class-specific metric\n",
    "    for i in range(y_pred.shape[1]):\n",
    "        y_pred_class = y_pred[:, i]\n",
    "        y_true_class = y_true[:, i]\n",
    "        true_positives = K.sum(K.round(K.clip(y_true_class * y_pred_class, 0, 1)))\n",
    "        possible_positives = K.sum(K.round(K.clip(y_true_class, 0, 1)))\n",
    "        recall = true_positives / (possible_positives + K.epsilon())\n",
    "        recall_by_class = recall_by_class + recall\n",
    "    return recall_by_class / y_pred.shape[1]\n",
    "\n",
    "#functions to find precision\n",
    "def balanced_precision(y_true, y_pred):\n",
    "    \"\"\"This function calculates the balanced precision metric\n",
    "    precision = TP / (TP + FP)\n",
    "    \"\"\"\n",
    "    precision_by_class = 0\n",
    "    # iterate over each predicted class to get class-specific metric\n",
    "    for i in range(y_pred.shape[1]):\n",
    "        y_pred_class = y_pred[:, i]\n",
    "        y_true_class = y_true[:, i]\n",
    "        true_positives = K.sum(K.round(K.clip(y_true_class * y_pred_class, 0, 1)))\n",
    "        predicted_positives = K.sum(K.round(K.clip(y_pred_class, 0, 1)))\n",
    "        precision = true_positives / (predicted_positives + K.epsilon())\n",
    "        precision_by_class = precision_by_class + precision\n",
    "    # return average balanced metric for each class\n",
    "    return precision_by_class / y_pred.shape[1]\n",
    "\n",
    "#functions to find f1 score\n",
    "def balanced_f1_score(y_true, y_pred):\n",
    "    \"\"\"This function calculates the F1 score metric\"\"\"\n",
    "    precision = balanced_precision(y_true, y_pred)\n",
    "    recall = balanced_recall(y_true, y_pred)\n",
    "    return 2 * ((precision * recall) / (precision + recall + K.epsilon()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e49f522-af83-402e-9cd1-e98bc7899da4",
   "metadata": {},
   "source": [
    "#### Define Model\n",
    "- preporcessor & encoder layers\n",
    "- dropout & dense layer with softmax activation function\n",
    "- output space dimension = # of classes "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "4f8a4ac9-7c39-45ed-8893-4fc3647bb1b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "#define model\n",
    "i = tf.keras.layers.Input(shape=(), dtype=tf.string, name='text')\n",
    "x = preprocessor(i)\n",
    "x = encoder(x)\n",
    "x = tf.keras.layers.Dropout(0.2, name=\"dropout\")(x['pooled_output'])\n",
    "x = tf.keras.layers.Dense(num_classes, activation='softmax', name=\"output\")(x)\n",
    "\n",
    "bert_model = tf.keras.Model(i, x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "5424bd9d-2a16-4f8f-bd8f-9b246925d448",
   "metadata": {},
   "outputs": [],
   "source": [
    "#model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "096ea57a-5a67-4433-8b85-bcc127ce4be2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# saving and loading the .h5 model\n",
    "# save model\n",
    "#model.save('Model.h5')\n",
    "#print('Model Saved!')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "82b53280-b5e8-47e8-bdbd-d24e8f3fee41",
   "metadata": {},
   "source": [
    "#### Compile & Fit Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "691b5eda-4e9e-45f8-9401-fa778597d8fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "2/2 [==============================] - 20s 7s/step - loss: 0.8202 - accuracy: 0.3800 - true_negatives_5: 19.0000 - false_positives_1: 31.0000 - auc: 0.3026 - balanced_recall: 0.3844 - balanced_precision: 0.3774 - balanced_f1_score: 0.3808 - val_loss: 0.8250 - val_accuracy: 0.5000 - val_true_negatives_5: 11.0000 - val_false_positives_1: 11.0000 - val_auc: 0.4845 - val_balanced_recall: 0.3958 - val_balanced_precision: 0.4048 - val_balanced_f1_score: 0.4002\n",
      "Epoch 2/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.7455 - accuracy: 0.5000 - true_negatives_5: 25.0000 - false_positives_1: 25.0000 - auc: 0.4976 - balanced_recall: 0.5156 - balanced_precision: 0.5156 - balanced_f1_score: 0.5156 - val_loss: 0.7868 - val_accuracy: 0.5000 - val_true_negatives_5: 11.0000 - val_false_positives_1: 11.0000 - val_auc: 0.5382 - val_balanced_recall: 0.3958 - val_balanced_precision: 0.4048 - val_balanced_f1_score: 0.4002\n",
      "Epoch 3/50\n",
      "2/2 [==============================] - 8s 4s/step - loss: 0.7072 - accuracy: 0.5800 - true_negatives_5: 29.0000 - false_positives_1: 21.0000 - auc: 0.5712 - balanced_recall: 0.5906 - balanced_precision: 0.5995 - balanced_f1_score: 0.5950 - val_loss: 0.7512 - val_accuracy: 0.6364 - val_true_negatives_5: 14.0000 - val_false_positives_1: 8.0000 - val_auc: 0.5806 - val_balanced_recall: 0.5938 - val_balanced_precision: 0.5804 - val_balanced_f1_score: 0.5870\n",
      "Epoch 4/50\n",
      "2/2 [==============================] - 8s 5s/step - loss: 0.6296 - accuracy: 0.6600 - true_negatives_5: 33.0000 - false_positives_1: 17.0000 - auc: 0.7080 - balanced_recall: 0.6752 - balanced_precision: 0.6786 - balanced_f1_score: 0.6769 - val_loss: 0.7190 - val_accuracy: 0.6818 - val_true_negatives_5: 15.0000 - val_false_positives_1: 7.0000 - val_auc: 0.6147 - val_balanced_recall: 0.6771 - val_balanced_precision: 0.6453 - val_balanced_f1_score: 0.6608\n",
      "Epoch 5/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.5983 - accuracy: 0.6600 - true_negatives_5: 33.0000 - false_positives_1: 17.0000 - auc: 0.7512 - balanced_recall: 0.6348 - balanced_precision: 0.6611 - balanced_f1_score: 0.6473 - val_loss: 0.7030 - val_accuracy: 0.6818 - val_true_negatives_5: 15.0000 - val_false_positives_1: 7.0000 - val_auc: 0.6054 - val_balanced_recall: 0.6771 - val_balanced_precision: 0.6453 - val_balanced_f1_score: 0.6608\n",
      "Epoch 6/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.5688 - accuracy: 0.6600 - true_negatives_5: 33.0000 - false_positives_1: 17.0000 - auc: 0.8000 - balanced_recall: 0.6559 - balanced_precision: 0.6879 - balanced_f1_score: 0.6715 - val_loss: 0.6925 - val_accuracy: 0.6818 - val_true_negatives_5: 15.0000 - val_false_positives_1: 7.0000 - val_auc: 0.6178 - val_balanced_recall: 0.6771 - val_balanced_precision: 0.6453 - val_balanced_f1_score: 0.6608\n",
      "Epoch 7/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.5254 - accuracy: 0.8600 - true_negatives_5: 43.0000 - false_positives_1: 7.0000 - auc: 0.8760 - balanced_recall: 0.8566 - balanced_precision: 0.8420 - balanced_f1_score: 0.8493 - val_loss: 0.6661 - val_accuracy: 0.7273 - val_true_negatives_5: 16.0000 - val_false_positives_1: 6.0000 - val_auc: 0.6529 - val_balanced_recall: 0.7604 - val_balanced_precision: 0.7083 - val_balanced_f1_score: 0.7335\n",
      "Epoch 8/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.5041 - accuracy: 0.8000 - true_negatives_5: 40.0000 - false_positives_1: 10.0000 - auc: 0.8840 - balanced_recall: 0.8000 - balanced_precision: 0.7978 - balanced_f1_score: 0.7989 - val_loss: 0.6396 - val_accuracy: 0.7273 - val_true_negatives_5: 16.0000 - val_false_positives_1: 6.0000 - val_auc: 0.6880 - val_balanced_recall: 0.7604 - val_balanced_precision: 0.7083 - val_balanced_f1_score: 0.7335\n",
      "Epoch 9/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.4887 - accuracy: 0.8600 - true_negatives_5: 43.0000 - false_positives_1: 7.0000 - auc: 0.9380 - balanced_recall: 0.8559 - balanced_precision: 0.8580 - balanced_f1_score: 0.8569 - val_loss: 0.6023 - val_accuracy: 0.7273 - val_true_negatives_5: 16.0000 - val_false_positives_1: 6.0000 - val_auc: 0.7707 - val_balanced_recall: 0.7604 - val_balanced_precision: 0.7083 - val_balanced_f1_score: 0.7335\n",
      "Epoch 10/50\n",
      "2/2 [==============================] - 8s 5s/step - loss: 0.4427 - accuracy: 0.8800 - true_negatives_5: 44.0000 - false_positives_1: 6.0000 - auc: 0.9532 - balanced_recall: 0.8606 - balanced_precision: 0.8446 - balanced_f1_score: 0.8525 - val_loss: 0.5657 - val_accuracy: 0.7727 - val_true_negatives_5: 17.0000 - val_false_positives_1: 5.0000 - val_auc: 0.8347 - val_balanced_recall: 0.7917 - val_balanced_precision: 0.7393 - val_balanced_f1_score: 0.7646\n",
      "Epoch 11/50\n",
      "2/2 [==============================] - 8s 5s/step - loss: 0.4176 - accuracy: 0.9000 - true_negatives_5: 45.0000 - false_positives_1: 5.0000 - auc: 0.9632 - balanced_recall: 0.9155 - balanced_precision: 0.9097 - balanced_f1_score: 0.9126 - val_loss: 0.5433 - val_accuracy: 0.7727 - val_true_negatives_5: 17.0000 - val_false_positives_1: 5.0000 - val_auc: 0.8616 - val_balanced_recall: 0.7917 - val_balanced_precision: 0.7393 - val_balanced_f1_score: 0.7646\n",
      "Epoch 12/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.3941 - accuracy: 0.9000 - true_negatives_5: 45.0000 - false_positives_1: 5.0000 - auc: 0.9710 - balanced_recall: 0.8864 - balanced_precision: 0.8801 - balanced_f1_score: 0.8833 - val_loss: 0.5256 - val_accuracy: 0.7727 - val_true_negatives_5: 17.0000 - val_false_positives_1: 5.0000 - val_auc: 0.8709 - val_balanced_recall: 0.7917 - val_balanced_precision: 0.7393 - val_balanced_f1_score: 0.7646\n",
      "Epoch 13/50\n",
      "2/2 [==============================] - 8s 5s/step - loss: 0.3874 - accuracy: 0.8800 - true_negatives_5: 44.0000 - false_positives_1: 6.0000 - auc: 0.9628 - balanced_recall: 0.8919 - balanced_precision: 0.8686 - balanced_f1_score: 0.8800 - val_loss: 0.5131 - val_accuracy: 0.8182 - val_true_negatives_5: 18.0000 - val_false_positives_1: 4.0000 - val_auc: 0.8853 - val_balanced_recall: 0.8229 - val_balanced_precision: 0.7768 - val_balanced_f1_score: 0.7992\n",
      "Epoch 14/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.3650 - accuracy: 0.9200 - true_negatives_5: 46.0000 - false_positives_1: 4.0000 - auc: 0.9810 - balanced_recall: 0.9271 - balanced_precision: 0.9332 - balanced_f1_score: 0.9301 - val_loss: 0.4926 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.8946 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 15/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.3435 - accuracy: 0.9200 - true_negatives_5: 46.0000 - false_positives_1: 4.0000 - auc: 0.9852 - balanced_recall: 0.9219 - balanced_precision: 0.9312 - balanced_f1_score: 0.9265 - val_loss: 0.4770 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9008 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 16/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.3524 - accuracy: 0.9200 - true_negatives_5: 46.0000 - false_positives_1: 4.0000 - auc: 0.9784 - balanced_recall: 0.9191 - balanced_precision: 0.9317 - balanced_f1_score: 0.9254 - val_loss: 0.4612 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9091 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 17/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.3287 - accuracy: 0.9200 - true_negatives_5: 46.0000 - false_positives_1: 4.0000 - auc: 0.9852 - balanced_recall: 0.9261 - balanced_precision: 0.9281 - balanced_f1_score: 0.9271 - val_loss: 0.4467 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9236 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 18/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2857 - accuracy: 0.9400 - true_negatives_5: 47.0000 - false_positives_1: 3.0000 - auc: 0.9962 - balanced_recall: 0.9375 - balanced_precision: 0.9495 - balanced_f1_score: 0.9435 - val_loss: 0.4336 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9298 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 19/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2924 - accuracy: 0.9200 - true_negatives_5: 46.0000 - false_positives_1: 4.0000 - auc: 0.9888 - balanced_recall: 0.9261 - balanced_precision: 0.9281 - balanced_f1_score: 0.9271 - val_loss: 0.4207 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9298 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 20/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2674 - accuracy: 0.9600 - true_negatives_5: 48.0000 - false_positives_1: 2.0000 - auc: 0.9928 - balanced_recall: 0.9688 - balanced_precision: 0.9722 - balanced_f1_score: 0.9705 - val_loss: 0.4104 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9298 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 21/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2531 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.4015 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9442 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 22/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2589 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9988 - balanced_recall: 0.9722 - balanced_precision: 0.9750 - balanced_f1_score: 0.9736 - val_loss: 0.3913 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9504 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 23/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2418 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9996 - balanced_recall: 0.9833 - balanced_precision: 0.9861 - balanced_f1_score: 0.9847 - val_loss: 0.3850 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9504 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 24/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2348 - accuracy: 0.9600 - true_negatives_5: 48.0000 - false_positives_1: 2.0000 - auc: 0.9968 - balanced_recall: 0.9686 - balanced_precision: 0.9686 - balanced_f1_score: 0.9686 - val_loss: 0.3785 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9504 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 25/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2320 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9984 - balanced_recall: 0.9853 - balanced_precision: 0.9844 - balanced_f1_score: 0.9848 - val_loss: 0.3713 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9504 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 26/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2312 - accuracy: 0.9400 - true_negatives_5: 47.0000 - false_positives_1: 3.0000 - auc: 0.9948 - balanced_recall: 0.9408 - balanced_precision: 0.9436 - balanced_f1_score: 0.9422 - val_loss: 0.3647 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9525 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 27/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.2117 - accuracy: 0.9600 - true_negatives_5: 48.0000 - false_positives_1: 2.0000 - auc: 0.9978 - balanced_recall: 0.9706 - balanced_precision: 0.9706 - balanced_f1_score: 0.9706 - val_loss: 0.3584 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9545 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 28/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1866 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3512 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9545 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 29/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1861 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9980 - balanced_recall: 0.9821 - balanced_precision: 0.9868 - balanced_f1_score: 0.9845 - val_loss: 0.3423 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9504 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 30/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1774 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3362 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9545 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 31/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1870 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3302 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9545 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 32/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1668 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3265 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9587 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 33/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1614 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9996 - balanced_recall: 0.9853 - balanced_precision: 0.9844 - balanced_f1_score: 0.9848 - val_loss: 0.3224 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9587 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 34/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1723 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3171 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9607 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 35/50\n",
      "2/2 [==============================] - 7s 5s/step - loss: 0.1632 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9996 - balanced_recall: 0.9844 - balanced_precision: 0.9853 - balanced_f1_score: 0.9848 - val_loss: 0.3132 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9628 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 36/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1553 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3103 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9628 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 37/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1629 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9988 - balanced_recall: 0.9833 - balanced_precision: 0.9861 - balanced_f1_score: 0.9847 - val_loss: 0.3082 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9607 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 38/50\n",
      "2/2 [==============================] - 8s 4s/step - loss: 0.1621 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3059 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9669 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 39/50\n",
      "2/2 [==============================] - 8s 4s/step - loss: 0.1350 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9996 - balanced_recall: 0.9844 - balanced_precision: 0.9853 - balanced_f1_score: 0.9848 - val_loss: 0.3041 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9607 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 40/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1254 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.3021 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9649 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 41/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1371 - accuracy: 0.9600 - true_negatives_5: 48.0000 - false_positives_1: 2.0000 - auc: 0.9986 - balanced_recall: 0.9667 - balanced_precision: 0.9737 - balanced_f1_score: 0.9701 - val_loss: 0.2987 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9649 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 42/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1467 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2943 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9711 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 43/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1375 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2881 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9752 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 44/50\n",
      "2/2 [==============================] - 8s 5s/step - loss: 0.1183 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2838 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9793 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 45/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1370 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9996 - balanced_recall: 0.9773 - balanced_precision: 0.9688 - balanced_f1_score: 0.9730 - val_loss: 0.2780 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9793 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 46/50\n",
      "2/2 [==============================] - 8s 5s/step - loss: 0.1310 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2751 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9793 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 47/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1161 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2731 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9793 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 48/50\n",
      "2/2 [==============================] - 7s 5s/step - loss: 0.1137 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2725 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9793 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 49/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1265 - accuracy: 0.9800 - true_negatives_5: 49.0000 - false_positives_1: 1.0000 - auc: 0.9996 - balanced_recall: 0.9821 - balanced_precision: 0.9868 - balanced_f1_score: 0.9845 - val_loss: 0.2723 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9793 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n",
      "Epoch 50/50\n",
      "2/2 [==============================] - 7s 4s/step - loss: 0.1256 - accuracy: 1.0000 - true_negatives_5: 50.0000 - false_positives_1: 0.0000e+00 - auc: 1.0000 - balanced_recall: 1.0000 - balanced_precision: 1.0000 - balanced_f1_score: 1.0000 - val_loss: 0.2721 - val_accuracy: 0.9091 - val_true_negatives_5: 20.0000 - val_false_positives_1: 2.0000 - val_auc: 0.9773 - val_balanced_recall: 0.8854 - val_balanced_precision: 0.8854 - val_balanced_f1_score: 0.8854\n"
     ]
    }
   ],
   "source": [
    "#define number of epochs\n",
    "n_epochs = 50\n",
    "\n",
    "METRICS = [\n",
    "      tf.keras.metrics.CategoricalAccuracy(name=\"accuracy\"),\n",
    "      tf.keras.metrics.TrueNegatives(),\n",
    "      tf.keras.metrics.FalsePositives(),\n",
    "      tf.keras.metrics.AUC(),\n",
    "      balanced_recall,\n",
    "      balanced_precision,\n",
    "      balanced_f1_score\n",
    "]\n",
    "\n",
    "#EarlyStopping callback to monitor validation loss\n",
    "#if metric doesn't improve for at least 3 epochs (patience = 3)\n",
    "    #training is interrupted and weights from epoch where the validation loss \n",
    "    #showed the best value (i.e. lowest) are restored (restore_best_weights = True)\n",
    "earlystop_callback = tf.keras.callbacks.EarlyStopping(monitor = \"val_loss\", \n",
    "                                                      patience = 3,\n",
    "                                                      restore_best_weights = True)\n",
    "#compile model\n",
    "bert_model.compile(optimizer = \"adam\",\n",
    "              loss = \"categorical_crossentropy\",\n",
    "              metrics = METRICS)\n",
    "#fit model\n",
    "bert_model = bert_model.fit(x_train, \n",
    "                      y_train, \n",
    "                      epochs = n_epochs,\n",
    "                      validation_data = (x_test, y_test),\n",
    "                      callbacks = [earlystop_callback], verbose=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "77a837c6-89ee-4cc6-bd2b-1546f4f0d5f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "#bert_model.history['accuracy']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "44a663dd-86d6-41e5-b8e7-e28cb40a55ec",
   "metadata": {},
   "source": [
    "### Plot Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "2751eb93-d44c-40bd-a287-f3362889926c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 3000x500 with 8 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = list(range(1, n_epochs+1))\n",
    "metric_list = list(bert_model.history.keys())\n",
    "num_metrics = int(len(metric_list)/2)\n",
    "\n",
    "fig, ax = plt.subplots(nrows=1, ncols=num_metrics, figsize=(30, 5))\n",
    "\n",
    "for i in range(0, num_metrics):\n",
    "  ax[i].plot(x, bert_model.history[metric_list[i]], marker=\"o\", label=metric_list[i].replace(\"_\", \" \"))\n",
    "  ax[i].plot(x, bert_model.history[metric_list[i+num_metrics]], marker=\"o\", label=metric_list[i+num_metrics].replace(\"_\", \" \"))\n",
    "  ax[i].set_xlabel(\"epochs\",fontsize=14)\n",
    "  ax[i].set_title(metric_list[i].replace(\"_\", \" \"),fontsize=20)\n",
    "  ax[i].legend(loc=\"lower left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "6aabeb10-3a74-46ae-a63f-3d199e9c12d7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['loss',\n",
       " 'accuracy',\n",
       " 'true_negatives_5',\n",
       " 'false_positives_1',\n",
       " 'auc',\n",
       " 'balanced_recall',\n",
       " 'balanced_precision',\n",
       " 'balanced_f1_score',\n",
       " 'val_loss',\n",
       " 'val_accuracy',\n",
       " 'val_true_negatives_5',\n",
       " 'val_false_positives_1',\n",
       " 'val_auc',\n",
       " 'val_balanced_recall',\n",
       " 'val_balanced_precision',\n",
       " 'val_balanced_f1_score']"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metric_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "62414bd5-4592-4d5e-83c7-7e369e2371d4",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}