Volume: 19 Issue-01 (January-June) 2024

---

EXPLORING COVID-19 VACCINE SENTIMENTS AND ADVERSE DRUG REACTIONS THROUGH ADVANCED MACHINE LEARNING AND DEEP LEARNING ANALYSIS

ABSTRACT

This research delves into a comprehensive analysis of public sentiments and adverse drug reactions (ADRs) associated with COVID-19 vaccines and hydroxychloroquine, leveraging data collected from Twitter and Google Form responses. The research adopts a structured methodology involving data collection, preprocessing, feature extraction, and evaluation using a combination of machine learning (ML) and deep learning models. In the preprocessing phase, techniques such as the removal of user names, punctuation, links, and stop words are applied to ensure consistency, with text converted to lowercase. Feature extraction methods, particularly N-grams-based approaches, are utilized to extract relevant messages from the preprocessed Twitter data. ML algorithms, with a specific focus on "Tri-grams with Q-SVM," are then assessed for predicting ADRs associated with COVISHIELD. Simultaneously, deep learning models, including LSTM, Bi-LSTM, CNN, and VAE-GANs, are employed to analyze sentiments surrounding COVID-19 vaccinations. The analysis culminates by underscoring the accuracy of "Tri-grams with Q-SVM" for ADR prediction and highlighting the efficacy of the VAE-GANs model in sentiment analysis. The abstract concludes by discussing the implications of the findings for policymakers and healthcare professionals, emphasizing the importance of accurate sentiment analysis in gauging public opinions. Additionally, it suggests future directions for enhanced vaccination campaigns and public health interventions.

FULL TEXT

PDF