ExpressAI

Authors: Samay Gupta, Rishee Mulchandani, Riti Dodiya

Abstract: In natural language processing, sentiment analysis has grown in importance, especially for multilingual and code- mixed languages like Hinglish. Analyzing sentiments in movie subtitles is still largely unexplored, despite themajority of sentiment analysis research concentrating on social media and product evaluations. This work introduces a newmethod for sentiment analysis of movie subtitles that combines deep learning-based sentiment classification, Hinglish-to-Hindi transliteration, and optical character recognition (OCR)-based subtitle extraction. The retrieved subtitles go through a transformer-based model for emotion analysis, preprocessing to eliminate noise, and transliteration into Hindi using the Google Translator API. Accurate sentiment classification is made possible by the suggested methodology, which captures the emotional tone of subtitles. Results from experiments show how well our method works with transliterated, noisy, and code- mixed text. By providing insights into the emotional dynamics of cinematic storylines, this research helps close the gap between sentiment analysis and multimedia content understanding. As a result, the words sentiment analysis and text analysis developed their paths to becoming important computational linguistics and text analysis components. [1

DOI: http://doi.org/