This study focuses on analyzing sentiment in mixed-indic social media text using aspect-based sentiment analysis. With the rise of social media, understanding user sentiment has become crucial for various applications such as brand monitoring, customer feedback analysis, and public opinion tracking. However, social media text often consists of mixed languages, where multiple languages or dialects are used within a single message. This mixed-indic language poses challenges for sentiment analysis techniques designed for monolingual or predominantly English text. To address this challenge, we propose a novel approach that leverages aspect-based sentiment analysis, which aims to identify and analyze sentiment towards specific aspects or entities mentioned in the text. By considering the sentiment at the aspect level, we can gain a deeper understanding of user opinions beyond the overall sentiment polarity. Our approach involves preprocessing techniques such as language identification, code-switching detection, and transliteration normalization to effectively handle the diverse linguistic elements present in mixed-indic text. We then employ state-of-the-art sentiment analysis algorithms, adapted and trained on mixed-indic corpora, to capture the sentiment associated with each aspect.

This study focuses on analyzing sentiment in mixed-indic social media text using aspect-based sentiment analysis.