Political Opinion Mining from Twitter

Abstract

Twitter is one of the most popular micro-blogging platform for people to express their political views in and around the elections. Hence during pre-elections twitter becomes a rich resource of data to understand the changing tenor of political leaders with time. During this time, when views, opinions and judgments are shared so prolifically through online media, tools which can provide the crux of this content are paramount. In this paper the authors have developed one such sentiment analysis tool to analyze the changing political views of persons with time. Using the tool they classify the tweets as positive, negative or neutral and studying it over time the authors successfully estimate the mood of the person. The authors have also developed a specialized phonetic dictionary to provide better approximation for most commonly used slangs and abbreviations.