• Version
• Download 16
• File Size 440.35 KB
• File Count 1
• Create Date 26 July 2025
• Last Updated 26 July 2025

    IMDB Sentiment Analysis Based on Comment by Using Machine Learning

MAMIDI TARANI, KURMAPU DURGAPRASAD.

       Assistant Professor, MCA Final Semester, Master of Computer Applications, Sanketika Vidya Parishad Engineering College, Vishakhapatnam, Andhra Pradesh, India.

Abstract:

            Sentiment analysis plays a significant role in understanding user opinions, product reviews, and market feedback by classifying textual data into sentiment categories. This project presents a lightweight, web-based sentiment analysis application using the Flask framework integrated with a Multinomial Naive Bayes classifier. The system is designed to classify movie reviews as either positive or negative, facilitating quick and accessible feedback analysis for users. Text data is pre-processed and transformed into numerical vectors using CountVectorizer, allowing the Naive Bayes model to perform effective feature-based classification. The classifier is trained on a small, clean dataset of movie reviews and demonstrates how simple yet powerful models can be applied in real-world sentiment classification tasks. To enable real-time sentiment predictions, the trained model and vectorizer are serialized using pickle and loaded within the Flask application. Users can submit their movie reviews through a simple HTML interface, and the system predicts and displays the sentiment classification immediately. Additionally, the application integrates Swagger documentation using Flasgger, providing clear API endpoint testing and making it easier for developers to extend or test the system. The modular structure of the application, including data processing, model training, and web deployment, makes it an effective learning tool for beginners in machine learning deployment workflows. This project demonstrates the end-to-end pipeline of building, training, and deploying a machine learning model using Python’s Flask micro-framework, emphasizing the practical application of sentiment analysis in web services. It highlights how sentiment analysis can be used in review systems, customer feedback monitoring, and content moderation. The system can be scaled further with larger datasets and additional NLP preprocessing for enhanced accuracy in practical applications. Overall, this project provides a hands-on implementation of a sentiment analysis system that is accessible, interpretable, and ready for integration into real-world environments.

Key Words: Sentiment Analysis, Naive Bayes, Flask, Natural Language Processing, Machine Learning, CountVectorizer