Twitter Sentiment Analysis Using Machine Learning Techniques

Abstract

This paper presents an effective sentiment analysis system designed to classify the polarity of tweets into positive, negative, or neutral sentiments. The framework utilizes supervised machine learning algorithms, including Logistic Regression, Support Vector Machines (SVM), and Random Forest, trained on the Sentiment140 dataset. Text preprocessing techniques such as tokenization, stopword removal, stemming, and TF-IDF vectorization are applied to improve classification performance. The proposed system achieves an accuracy of 87.2% with SVM, outperforming other baseline models. This solution offers scalable deployment in social media monitoring, political campaign tracking, and customer feedback analysis.