Publication

Abstract : Abstract: Social media platforms, such as Twitter, allow users to share their thoughts and opinions on various topics, including pandemics like COVID-19. This data can be used to analyze public sentiment and assess the severity of different Corona variants. In this study, a new framework called SENSECOR is proposed to perform opinion mining on Twitter data. SENSECOR uses natural language processing techniques to identify the severity levels and most common symptoms associated with Corona variants. The dataset includes over 160,000 tweets related to COVID-19. SENSECOR is evaluated against several deep learning models, including RoBERTa, BERT, ELECTRA, XLNet, LSTM, and BiLSTM, as well as traditional machine learning models. The results show that SENSECOR achieves the highest accuracy rate of 91%, surpassing all other methods. This suggests that SENSECOR is a promising tool for assessing the severity of Corona variants and identifying the most common associated symptoms. Graphical abstract: Graphical abstract of SENSECOR framework [Figure not available: see fulltext.] © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.