Effectiveness of GNN based approach for Topic Classification of Telugu Text

Telugu, a Dravidian language spoken by over 80 million people, presents unique difficulties for text processing algorithms. Unlike English, it possesses complex morphology, agglutinative word formation, and rich context-dependent meanings, making it challenging for traditional NLP techniques. This abstract explores these challenges and argues why directly processing Telugu, rather than relying on English translation, is crucial for several reasons such as loss of nuancing and context, semantic ambiguities and idiomatic expressions can be misinterpreted, leading to inaccurate results in tasks like sentiment analysis or question answering. Further, we have utilized various state-of-art techniques to process the text directly in Telugu language rather than converting the text into English. Processing Telugu text directly allows for deeper analysis of its linguistic structure, literary works, and cultural expressions. This fosters a better understanding of Telugu heritage and facilitates its preservation for future generations.