Sentiment Analysis of IMDB Movie Reviews Using Logistic Regression and NLTK

Sentiment analysis, a subfield of Natural Language Processing (NLP), focuses on identifying emotional polarity in text and has become essential for understanding public opinion across digital platforms. This paper presents a sentiment classification model developed using the IMDB movie review dataset containing 50,000 labeled samples. The methodology employs text preprocessing techniques including tokenization, stopword removal, and stemming, followed by TF-IDF vectorization for feature extraction. A Logistic Regression classifier was trained for binary sentiment classification, achieving 88.7% accuracy on the test set. The model was deployed through a Flask-based web interface enabling real-time sentiment predictions. This work demonstrates that traditional machine learning approaches, when combined with appropriate preprocessing and feature engineering, can deliver reliable performance for sentiment analysis tasks.