Automated Sentiment Analysis of South Indian Movie Reviews Using Machine Learning

Sentiment Analysis (SA) has traditionally relied on manual processing of reviews, which is impractical due to the vast amount of data and subjective nature of human interpretation. This study automates the classification of movie reviews using different machine learning techniques with the focus on South Indian films by analyzing the IMDb dataset. Feature extraction is implemented through term frequencyinverse document frequency (TF-IDF) to represent the text data effectively. This paper evaluates the performance of several algorithms like Logistic Regression, Random Forest, Linear Support Vector Classifier, Multinomial Naive Bayes, Decision Tree, Extra Trees Classifier and K-Nearest Neighbors Classifier. Among these methods, the Multinomial NB model proved to be the most effective by achieving an accuracy of 86.71 %, precision of 86.71 %, recall of 86.69 % and F1-score of 86.69 %. Integrating this model into a Flask-based web application supported by a SQLite database enables users to submit and analyze movie reviews efficiently. In addition to simplifying the evaluation procedure, this automated method gives business experts useful insights on audience sentiment, thereby enhancing the decision-making process and improving feedback mechanisms for South Indian films.