Prediction of election result by enhanced sentiment analysis on twitter data using classifier ensemble Approach

Abstract

Sentiment analysis is the computational study of opinions, sentiments, evaluations, attitudes, views and emotions expressed in text. It refers to a classification problem where the main focus is to predict the polarity of words and then classify them into positive or negative sentiment. Sentiment analysis over Twitter offers people a fast and effective way to measure the public's feelings towards their party and politicians. The primary issue in previous sentiment analysis techniques is the determination of the most appropriate classifier for a given classification problem. If one classifier is chosen from the available classifiers, then there is no surety in the best performance on unseen data. So to reduce the risk of selecting an inappropriate classifier, we are combining the outputs of a set of classifiers. Thus in this paper, we use an approach that automatically classifies the sentiment of tweets by combining machine learning classifiers with lexicon based classifier. The new combination of classifiers are SentiWordNet classifier, naive bayes classifier and hidden markov model classifier. Here positivity or negativity of each tweet is determined by using the majority voting principle on the result of these three classifiers. Thus we were used this sentiment classifier for finding political sentiment from real time tweets. Thus we have got an improved accuracy in sentiment analysis using classifier ensemble Approach. Our method also uses negation handling and word sense disambiguation to achieve high accuracy.