Real-Time Social Media Sentiment Analysis: A Comparative Study of LightGBM and SVM

Abstract

This paper explores the effectiveness of using LightGBM for sentiment analysis on Twitter, comparing it to Support Vector Machine (SVM), a traditional machine learning approach. LightGBM's ability to handle high-dimensional data and quick processing speed makes it an ideal choice for real-time monitoring of social media sentiment. The study utilizes the Sentiment140 dataset, which contains over 1.6 million tweets, and compares the performance of LightGBM and SVM based on accuracy, precision, recall, and F1 scores. Experimental results reveal that LightGBM achieves superior performance with an accuracy of 91%, proving it to be more efficient for real-time sentiment analysis tasks.

Topics & Concepts