This repository contains code and resources for sentiment analysis on Tamil Code-Mixed Data using various BERT (Bidirectional Encoder Representations from Transformers) variants. The primary objective is to leverage pre-trained BERT models for accurate sentiment classification in code-mixed Tamil text.
Sentiment analysis in multilingual contexts like code-mixed Tamil text presents challenges due to language variations and informal expressions. This project focuses on utilizing BERT-based models, known for their contextual understanding, to capture sentiment in code-mixed Tamil language data.
The dataset utilized in this project comprises code-mixed Tamil texts extracted from social media, online forums, or conversational data. It includes a mix of Tamil and another language, posing challenges for sentiment classification.