Development of knowledge based sentiment analysis system using lexicon approach on twitter data

Abstract

In today's digital world, knowledge management systems aim at extracting semantic and knowledgeable information by examining natural language impressions. Social networking sites such as Twitter are becoming an integral part of our daily lives, contributing to a large increase in the growth of knowledgeable data. It is one of the sources of information especially for detecting interest of people. So, in this paper we propose to use a lexicon based approach to analyse knowledge based sentiment from tweets. We use a set of words with pre-defined polarity model to pick out polarity words with semantic scores assigned to them by classifying them using Naïve Bayes classification, and incorporates part of speech tagging. We then incorporate Natural Language Processing techniques on the resulting score to further improve the accuracy. We are also proposing additional knowledgeable sentiment detection using features like emoticons to assess sentiment. The proposed model achieved a minimal error rate of 0.36 and accuracy of 97% when compared to other similar methods.