Sentiment Analysis on Social Media Opinions: A Survey of Machine Learning and Lexicon-Based Approaches

Social media platforms, particularly Twitter, have become a rich source of public opinion, influencing various sectors such as politics, business, and social movements. Sentiment analysis plays a crucial role in interpreting and classifying emotions embedded in user-generated content. This paper provides a comprehensive survey of sentiment analysis methodologies, comparing machine learning-based and lexicon-based approaches. The study explores different frameworks, techniques, and challenges associated with sentiment classification, including sarcasm detection, data noise, and domain dependency. The paper reviews key contributions from recent studies, such as Qi et al. (2023), Karim et al., Dong & Lian (2021), Maqsood et al. (2020), Niu et al. (2021), and Nedjah et al. (2022), highlighting advancements and research gaps. Finally, the study identifies future research directions for improving sentiment analysis models by integrating deep learning, multimodal approaches, and real-time sentiment tracking.