AI-Driven Models for Combating Public Health Misinformation

 This research investigates the critical challenges of health-related misinformation and sentiment analysis on social media platforms, with a particular focus on content related to COVID-19 and vaccines. The study employs a comprehensive methodology combining natural language processing techniques, machine learning algorithms, and sentiment analysis to detect and classify health-related content. Through extensive data preprocessing, feature engineering using TF-IDF vectorization, and n-gram analysis, the research analyzed a dataset comprising 1,999 tweets categorized into positive, negative, and neutral sentiments. Multiple machine learning models were implemented, with a Voting Classifier achieving the highest accuracy of 89%, followed by Logistic Regression (88.60%), Random Forest (88.40%), and Gradient Boosting (86.00%). The study addresses the significant challenge of class imbalance in the dataset, consisting of 1,761 neutral, 195 positive, and 43 negative tweets. Key findings reveal the effectiveness of combined machine learning approaches in detecting health misinformation, with recent studies showing up to 91% accuracy in identifying false health claims and 87% precision in source credibility assessment. The research also highlights the distinction between disinformation and misinformation in health contexts and their propagation patterns on social media platforms. These findings contribute to the process of more efficient strategies for combating health misinformation while maintaining accurate sentiment analysis in public health communications.