Incremental Real-time Learning Framework for Sentiment Classification: Indian General Election 2019, A Case Study
Sharmistha P. Chatterjee, Sushmita Gupta · 2021
Indian General Election 2019 was one of the major global political events. The prominence of social media in contemporary life, and the ubiquity of political messaging on it has necessitated a systematic study of sentiments and make inferences about future moods and trends. In absence of standard tools or software we propose a machine learning-based generic system framework and REST (Representation State Transfer) plugin component, that extracts and filters authentic tweets from Twitter, captures the prevalent mood, and predicts the sentiment of any live incoming tweet in a resource-constrained setup. The system can predict user sentiments by discovering completely new features from the web, along with the process of continuous incremental learning and improvement of model accuracy.