Improved Twitter Sentiment Analysis Using N Gram Feature Selection and Combinations

Abstract: Sentiment analysis is an important research area that identifies the people�s sentiment underlying a text. Sentiment analysis widely studied in data mining. Sentiment analyses of tweets are widely studied. After reviewing and studying the current research on sentiment analysis, the goal of the proposed method is to get the more effective results of sentiment analysis on tweets. The aim of this paper is to improve the performance to classify the tweets with sentiment information. We use a feature combination scheme which uses the sentiment lexicons and extracted tweets n gram of high performance gain. We evaluate the performance of three popular machine learning classifiers among which Kern lab classifier achieves the highest accuracy rate.

Keywords: Data Mining; Sentiment Analysis; Twitter; Classification; Supervised learning Ngram; Feature Selection; Sentiment Lexicon.