Aspect Based Sentiment Analysis on Codemix Languages

Abstract In the era of social media, where multilingual conversations are prevalent, analyzing code-mixed text poses unique challenges. This project presents a comparative analysis of sentiment analysis and aspect-based sentiment analysis on code-mixed data using advanced techniques like Large Language Models (LLM), BERT, and Naive Bayes. Sentiment analysis categorizes text into positive, negative, or neutral sentiments, while aspect-based analysis identifies opinions on specific topics, such as "price" or "quality" in reviews. By focusing on code-mixed text, this study compares the effectiveness of each method in understanding sentiments and specific opinions, paving the way for improved applications in multilingual settings.