Sentiment Analysis on Social Media Using Transformer Models

AbstractThis paper explores the application of transformer models for sentiment analysis on social media platforms. Social mediapresents unique challenges for sentiment classification due to informal language, slang, and sarcastic expressions.Transformer models like BERT, RoBERTa, and DistilBERT, through their contextual understanding and attentionmechanisms, outperform traditional methods in capturing sentiment from noisy text. The study outlines the methodologyof applying these models, including data preprocessing, model training, and evaluation. It also discusses the challengesposed by bias, computation cost, and ambiguity in social media language. The paper concludes by highlighting futuretrends, including multilingual sentiment analysis, few-shot learning, and model explainability, positioning transformers asthe future of real-time opinion mining.