SENTIMENT ANALYSIS OF PUBLIC HEALTH SOCIAL MEDIA COMMENT USING EXPERT ANNOTATION

Abstract

Sentiment analysis has become a critical tool for organizations and researchers to understand user sentiment. However, it faces challenges such as managing noisy data, interpreting sarcasm or irony, and adapting to the evolving nature of language, especially in public health where users often express opinions about their health conditions, healthcare experiences, and complex medical terminology on various topics. Addressing these challenges is crucial to maintaining the integrity of sentiment analysis results. Hence, this study analyzes public health social media user comments using a structured sentiment analysis framework. The framework includes dataset collection and annotation, text preprocessing, feature vectorization, and text classification. The results show the model achieved 98% accuracy, demonstrating strong predictive performance. This accuracy falls within the 'Excellent Classification' range which is widely recognized as an industry standard for high-performing models, indicating that the model is not only accurate but operates with a level of precision and recall that exceeds the benchmarks typically required in practical applications.