_AI_Project02 (1).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "ddb1201d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package vader_lexicon to\n",
      "[nltk_data]     C:\\Users\\yazga\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package vader_lexicon is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "\"@AaronKaviiri Well, it's not like we've a choice when they still contribute a huge chunk to our countries ' budgets, fund our essential services including education, military and health (donated Covid-19 vaccines), huge loans and grants! Worst of it, our leaders making home hell for us to stay\""
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from textblob import TextBlob\n",
    "import sys\n",
    "import twint\n",
    "import nest_asyncio\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import nltk\n",
    "nltk.download('vader_lexicon')\n",
    "\n",
    "import re\n",
    "import string\n",
    "from wordcloud import WordCloud, STOPWORDS\n",
    "from PIL import Image\n",
    "from nltk.sentiment.vader import SentimentIntensityAnalyzer\n",
    "from nltk.stem import SnowballStemmer\n",
    "from nltk.sentiment.vader import SentimentIntensityAnalyzer\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "\n",
    "df = pd.read_csv(\"/Users/yazga/Downloads/covid19_education_data.csv\")\n",
    "counts = df['tweet'].values \n",
    "\n",
    "df.insert(36, \"Sentiment\", 0, True)\n",
    "\n",
    "#df.drop(columns=['id', 'conversation_id','created_at','date','time','timezone',\n",
    "      #   'user_id','username','name','place','language','mentions','urls',\n",
    "       #  'photos','replies_count','retweets_count','likes_count','hashtags',\n",
    "       #  'cashtags','link','retweet','quote_url','video','thumbnail',\n",
    "       #  'near','geo','source','user_rt_id','user_rt','retweet_id',\n",
    "       #  'reply_to','retweet_date','translate','trans_src','trans_dest'], axis=1, inplace=True)\n",
    "\n",
    "\n",
    "counts[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c50314de",
   "metadata": {},
   "outputs": [],
   "source": [
    "def percentage(part,whole):\n",
    " return 100 * float(part)/float(whole)\n",
    " \n",
    "\n",
    "noOfTweet = 500\n",
    "\n",
    "\n",
    "positive = 0\n",
    "negative = 0\n",
    "neutral = 0\n",
    "polarity = 0\n",
    "tweet_list = []\n",
    "neutral_list = []\n",
    "negative_list = []\n",
    "positive_list = []\n",
    "\n",
    "\n",
    "\n",
    "for count in counts:\n",
    "  analysis = TextBlob(count)\n",
    "  score = SentimentIntensityAnalyzer().polarity_scores(count)\n",
    "  neg = score['neg']\n",
    "  neu = score['neu']\n",
    "  pos = score['pos']\n",
    "  comp = score['compound']\n",
    "  polarity += analysis.sentiment.polarity\n",
    "  \n",
    "  if neg > pos:\n",
    "    negative_list.append(count)\n",
    "    df.at[count,'Sentiment']=0\n",
    "    negative += 1\n",
    "  elif pos > neg:\n",
    "    positive_list.append(count)\n",
    "    df.at[count,'Sentiment']=2\n",
    "    positive += 1\n",
    "  \n",
    "  elif pos == neg:\n",
    "    neutral_list.append(count)\n",
    "    df.at[count,'Sentiment']=1\n",
    "    neutral += 1\n",
    "\n",
    "\n",
    "positive = percentage(positive, noOfTweet)\n",
    "negative = percentage(negative, noOfTweet)\n",
    "neutral = percentage(neutral, noOfTweet)\n",
    "polarity = percentage(polarity, noOfTweet)\n",
    "positive = format(positive, '.1f')\n",
    "negative = format(negative, '.1f')\n",
    "neutral = format(neutral, '.1f')\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7624e2de",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "\n",
    "tweets = df['tweet']\n",
    "sentiment = df['Sentiment']\n",
    "\n",
    "\n",
    "train_data, test_data, y_train, y_test=train_test_split(tweets,\n",
    "                                                  sentiment,\n",
    "                                                 test_size=0.30,\n",
    "                                                 random_state=42)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "feb67ed9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import string\n",
    "from nltk.tokenize import word_tokenize\n",
    "from nltk.corpus import stopwords\n",
    "\n",
    "train_all_data=list()\n",
    "lines = counts\n",
    "\n",
    "for line in lines:\n",
    "    tokens=word_tokenize(line)\n",
    "    tokens=[w.lower() for w in tokens]\n",
    "    table=str.maketrans('','',string.punctuation)\n",
    "    stripped = [w.translate(table) for w in tokens]\n",
    "    words=[word for word in stripped if word.isalpha()]\n",
    "    stop_words=set(stopwords.words('english'))\n",
    "    words=[w for w in words if not w in stop_words]\n",
    "    train_all_data.append(words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "24967720",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['aaronkaviiri',\n",
       " 'well',\n",
       " 'like',\n",
       " 'choice',\n",
       " 'still',\n",
       " 'contribute',\n",
       " 'huge',\n",
       " 'chunk',\n",
       " 'countries',\n",
       " 'budgets',\n",
       " 'fund',\n",
       " 'essential',\n",
       " 'services',\n",
       " 'including',\n",
       " 'education',\n",
       " 'military',\n",
       " 'health',\n",
       " 'donated',\n",
       " 'vaccines',\n",
       " 'huge',\n",
       " 'loans',\n",
       " 'grants',\n",
       " 'worst',\n",
       " 'leaders',\n",
       " 'making',\n",
       " 'home',\n",
       " 'hell',\n",
       " 'us',\n",
       " 'stay']"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_all_data[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "166a3668",
   "metadata": {
    "scrolled": true
   },
   "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>tweet</th>\n",
       "      <th>Sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>@AaronKaviiri Well, it's not like we've a choi...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>@jonlevyBU The fact that u have to defend your...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>@StevenDelDuca @OntLiberal Ok, I own 0 guns, a...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>@MicheleWojnar @DemocraticDaisy The pro death ...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Covid related lockdowns in 2020 added to our w...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>@Lis_A_Martinez @TedMielczarek @PamBelluck not just to avoid hospitalization or death, but to avoid Long COVID.   We first need studies showing which therapeutics actually work in this regard.  Unfortunately, two years into this pandemic, that has not yet been done.  First we catch up on the data and public education....</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>@fucksakings i jus cant imagine myself in secondary education in a covid world. with college, there more option to protect myself and the prof. but in hs, the district prolly doesnt give a shit so i wouldn’t wanna be there</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>@Awesome_Knight5 @jaywhite81 @politicalplayer @NateSilver538 Pretty sure they just wanted to go about their daily lives and take their chances with Covid. Seems stupid to try and completely eliminate a highly-contagious virus while simultaneously ignoring the negative impact restrictions have on mental health, drug abuse, education, etc.</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Took, like, 3 days for a new conspiracy theory about how m*nkeypox is actually a sh*ngles reaction to the COVID vaccine and that this whole thing is an elaborate cover-up.  This is what you get when you stop properly funding education — an easily-duped population of simpletons.</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Oregon’s COVID-19 case rate is rising again, and that has state health and education officials worried.  https://t.co/AzxZ3BlRVj</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5155 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                                                tweet  \\\n",
       "0                                                   @AaronKaviiri Well, it's not like we've a choi...   \n",
       "1                                                   @jonlevyBU The fact that u have to defend your...   \n",
       "2                                                   @StevenDelDuca @OntLiberal Ok, I own 0 guns, a...   \n",
       "3                                                   @MicheleWojnar @DemocraticDaisy The pro death ...   \n",
       "4                                                   Covid related lockdowns in 2020 added to our w...   \n",
       "...                                                                                               ...   \n",
       "@Lis_A_Martinez @TedMielczarek @PamBelluck not ...                                                NaN   \n",
       "@fucksakings i jus cant imagine myself in secon...                                                NaN   \n",
       "@Awesome_Knight5 @jaywhite81 @politicalplayer @...                                                NaN   \n",
       "Took, like, 3 days for a new conspiracy theory ...                                                NaN   \n",
       "Oregon’s COVID-19 case rate is rising again, an...                                                NaN   \n",
       "\n",
       "                                                    Sentiment  \n",
       "0                                                         0.0  \n",
       "1                                                         0.0  \n",
       "2                                                         0.0  \n",
       "3                                                         0.0  \n",
       "4                                                         0.0  \n",
       "...                                                       ...  \n",
       "@Lis_A_Martinez @TedMielczarek @PamBelluck not ...        0.0  \n",
       "@fucksakings i jus cant imagine myself in secon...        2.0  \n",
       "@Awesome_Knight5 @jaywhite81 @politicalplayer @...        0.0  \n",
       "Took, like, 3 days for a new conspiracy theory ...        0.0  \n",
       "Oregon’s COVID-19 case rate is rising again, an...        0.0  \n",
       "\n",
       "[5155 rows x 2 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e5ac54cd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Vocubulary size : 4805\n"
     ]
    }
   ],
   "source": [
    "import gensim\n",
    "from gensim.test.utils import common_texts\n",
    "from gensim.models import Word2Vec\n",
    "embedding_dim= 50\n",
    "model=gensim.models.Word2Vec(sentences=train_all_data,\n",
    "                            vector_size=embedding_dim,\n",
    "                            window=5,\n",
    "                            min_count=1,\n",
    "                            workers=4,)\n",
    "\n",
    "words= len(model.wv)\n",
    "print('Vocubulary size : %d' %(words))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b98fa7d8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('medical', 0.9985640645027161),\n",
       " ('actually', 0.9984803199768066),\n",
       " ('ill', 0.9984423518180847),\n",
       " ('better', 0.9984312653541565),\n",
       " ('fucking', 0.9983932375907898),\n",
       " ('kenney', 0.998374342918396),\n",
       " ('voting', 0.9983043670654297),\n",
       " ('income', 0.9982598423957825),\n",
       " ('making', 0.9982514381408691),\n",
       " ('problems', 0.9982492327690125)]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.wv.most_similar('grade')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a7d102ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename= 'embedding_Word2Vec.txt'\n",
    "model.wv.save_word2vec_format(filename, binary=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a256c922",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "\n",
    "embeddings_index={}\n",
    "f = open(os.path.join('','embedding_Word2Vec.txt'), encoding='utf-8')\n",
    "\n",
    "for line in f:\n",
    "    values=line.split()\n",
    "    word=values[0]\n",
    "    coefs=np.asarray(values[1:])\n",
    "    embeddings_index[word]=coefs\n",
    "f.close()    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "4fe52654",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique tokens:5693\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.text import Tokenizer\n",
    "\n",
    "train_data=train_data.astype(str)\n",
    "test_data=test_data.astype(str)\n",
    "tokenizer = Tokenizer()\n",
    "tokenizer.fit_on_texts(train_data)\n",
    "\n",
    "X_train = tokenizer.texts_to_sequences(train_data)\n",
    "X_test =  tokenizer.texts_to_sequences(test_data)\n",
    "\n",
    "word_index = tokenizer.word_index\n",
    "print('Unique tokens:%d' %len(word_index))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b361caa6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "@AaronKaviiri Well, it's not like we've a choice when they still contribute a huge chunk to our countries ' budgets, fund our essential services including education, military and health (donated Covid-19 vaccines), huge loans and grants! Worst of it, our leaders making home hell for us to stay\n",
      "[32]\n"
     ]
    }
   ],
   "source": [
    "print(train_data[0])\n",
    "print(X_train[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "422f4c2d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Maxlen: 97\n",
      "Training data: (3608, 97)\n",
      "Test data: (1547, 97)\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.sequence import pad_sequences\n",
    "\n",
    "maxlen = max([len(x) for x in X_train])\n",
    "print('Maxlen:',maxlen)\n",
    "\n",
    "X_train = pad_sequences(X_train, padding='post', maxlen=maxlen)\n",
    "X_test = pad_sequences(X_test, padding='post', maxlen=maxlen)\n",
    "\n",
    "print('Training data:', X_train.shape)\n",
    "print('Test data:', X_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5b4a8bdc",
   "metadata": {},
   "outputs": [],
   "source": [
    "num_words = len(word_index) + 1\n",
    "embedding_matrix = np.zeros((num_words, embedding_dim))\n",
    "\n",
    "for word, i in word_index.items():\n",
    "    if i>num_words:\n",
    "        continue\n",
    "    embedding_vector=embeddings_index.get(word)\n",
    "    if embedding_vector is not None:\n",
    "        embedding_matrix[i]=embedding_vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b165e44f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5694\n"
     ]
    }
   ],
   "source": [
    "print(num_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "416db181",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding (Embedding)        (None, 97, 50)            284700    \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d (Global (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 10)                510       \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 285,221\n",
      "Trainable params: 521\n",
      "Non-trainable params: 284,700\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import *\n",
    "from keras.layers.embeddings import Embedding\n",
    "from keras.initializers import Constant\n",
    "\n",
    "model=Sequential()\n",
    "model.add(Embedding(num_words,\n",
    "                   embedding_dim,\n",
    "                   embeddings_initializer=Constant(embedding_matrix),\n",
    "                   input_length=maxlen,\n",
    "                   trainable=False))\n",
    "model.add(GlobalMaxPool1D())\n",
    "model.add(Dense(10, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "db76a5ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "005d73db",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "113/113 [==============================] - 0s 2ms/step - loss: 0.5235 - accuracy: 0.6788 - val_loss: 0.1985 - val_accuracy: 0.9412\n",
      "Epoch 2/10\n",
      "113/113 [==============================] - 0s 876us/step - loss: 0.1471 - accuracy: 0.9374 - val_loss: 0.1158 - val_accuracy: 0.9412\n",
      "Epoch 3/10\n",
      "113/113 [==============================] - 0s 895us/step - loss: 0.1075 - accuracy: 0.9374 - val_loss: 0.0965 - val_accuracy: 0.9412\n",
      "Epoch 4/10\n",
      "113/113 [==============================] - 0s 903us/step - loss: 0.0932 - accuracy: 0.9374 - val_loss: 0.0868 - val_accuracy: 0.9412\n",
      "Epoch 5/10\n",
      "113/113 [==============================] - 0s 813us/step - loss: 0.0865 - accuracy: 0.9210 - val_loss: 0.0822 - val_accuracy: 0.8985\n",
      "Epoch 6/10\n",
      "113/113 [==============================] - 0s 851us/step - loss: 0.0824 - accuracy: 0.8961 - val_loss: 0.0788 - val_accuracy: 0.8985\n",
      "Epoch 7/10\n",
      "113/113 [==============================] - 0s 808us/step - loss: 0.0791 - accuracy: 0.8961 - val_loss: 0.0760 - val_accuracy: 0.8985\n",
      "Epoch 8/10\n",
      "113/113 [==============================] - 0s 652us/step - loss: 0.0761 - accuracy: 0.8961 - val_loss: 0.0733 - val_accuracy: 0.8985\n",
      "Epoch 9/10\n",
      "113/113 [==============================] - 0s 708us/step - loss: 0.0733 - accuracy: 0.8961 - val_loss: 0.0707 - val_accuracy: 0.8985\n",
      "Epoch 10/10\n",
      "113/113 [==============================] - 0s 704us/step - loss: 0.0707 - accuracy: 0.8961 - val_loss: 0.0684 - val_accuracy: 0.8985\n"
     ]
    }
   ],
   "source": [
    "model_history = model.fit(X_train,y_train, epochs=10, validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "d36fd8fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.8961\n",
      "Testing Accuracy:  0.8985\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model.evaluate(X_train, y_train, verbose=False)\n",
    "print('Training Accuracy:  {:.4f}'.format(accuracy))\n",
    "\n",
    "loss, accuracy = model.evaluate(X_test, y_test, verbose=False)\n",
    "print('Testing Accuracy:  {:.4f}'.format(accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "82cab8cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "acc = model_history.history['accuracy']\n",
    "val_acc = model_history.history['val_accuracy']\n",
    "loss = model_history.history['loss']\n",
    "val_loss = model_history.history['val_loss']\n",
    "\n",
    "plt.figure(figsize=(12,6))\n",
    "plt.subplot(1,2,1)\n",
    "plt.plot(acc, 'b', label='Training Accuracy')\n",
    "plt.plot(val_acc, 'r', label='Validation accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy Value')\n",
    "plt.title('Training and Validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "plt.plot(loss, 'b', label='Training loss')\n",
    "plt.plot(val_loss, 'r' ,label='Validation loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss Value')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2dd79823",
   "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": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}