Twitter Sentiment Analysis

• This is the repository for Twitter Sentiment Analysis Dataset uploaded on Kaggle https://www.kaggle.com/kazanova/sentiment140
• We did Data Analysis from different perspectives to know in-detail about the sentiments of tweets
• We defined labels in negative and positive tweets
• Then we cleaned the text and remove the url, numbers, punctuation and stopwords we defined.
• Then we split the dataset into training test set and tokenize words in tweets
• We trained using pretrained model and build our model with tensorflow.keras.

Model Building

• We are got the Validation Accuracy of 76.97%
• And Loss of 0.47.