An Efficient Expert System for Real time Sentiment Analysis Using GPU

Abstract

Sentiment Analysis has become a research hot-spot with the rapid development of social network websites. However, unannotated data, large-scale unstructured data, and low accuracy has plagued current approaches. Considering its significance in business intelligence and decision-making, there exists a need to develop robust expert systems for real-time sentiment analysis. Twitter is a social network application with millions of users expressing their sentiments. In this work, we explored comprehensively the methodologies applied in sentiment classification over Twitter data: rule-based, machine learning (ML) and deep learning (DL) methods. Our datasets, based on tweet topics ranging from general, news feed, airlines and politics are crawled and manually cleaned with the principle of Naturally Annotated Big Data. newline