Date of Publication :5th February 2025

Abstract: COVID-19 is a life changing incident for the whole world. Its variant Omicron has caused vivid disturbances and health issues for millions of people around the world. To get more idea about the people’s thoughts, there is a need for sentiment analysis module to help the government and health sectors for improvising to stop the spread of Omicron. Twitter is a high level micro blogging site to gather public’s views about Omicron. The tweets are gathered from twitter for the keyword Omicron, the se text data are labeled according to its sentiment using TextBlob algorithm. This labeled data undergoes various trainings using a few classifier models. The deep learning models like LSTM, BiLSTM, BERT and Roberta are trained for tweets dataset. Roberta outperforms due to its unique feature of dynamic masking and the usage of transformer models in it with an accuracy of 91%.

Reference :