Sentiment Analysis of Tweets During COVID-19 Pandemic Using BLSTM

social media is the new way for people to express and share their thoughts and it plays a huge role in spreading the anxiety widely during the pandemics and Twitter is one of the social media channels with high number of users and daily tweets. Clarifying and understanding what people thinks and their shared opinions during such hard times can reduce the burden on the health systems and entities and redirects the concerned entities by highlighting the areas to spread the awareness in. The aim of this study is to analyze and assess the sentiment of the Tweets shared during the COVID-19 pandemic and testing the Bidirectional Long Term Short Memory (BLTSM) of Recurrent Neural Network (RNN) in predicting the sentiment class which are Negative, Positive and Neutral. The results show slight difference between the positive and negative tweets which needs an attention to spread the awareness and hence, reduce the negativity. Furthermore, the BLSTM predicted the sentiment classes (Negative, Positive and Neutral) and obtained 86.15% accuracy rate. The high accuracy concludes that BLSTM can adapt and predict the sentiment of a text.