Opinion Mining and Sentiment Analysis on Twitter

Abstract

Twitter platform is valuable to follow the public sentiments. Knowing users point of views and reasons behind them at various point is an important study to take certain decisions. Categorization of positive and negative opinions is a process of sentiment analysis. It is very useful for people to find sentiment about the person, product etc. before they actually make opinion about them. In this paper Latent Dirichlet Allocation (LDA) based models are defined. Where the first model that is Foreground and Background LDA (FB-LDA) can remove background topics and selects foreground topics from tweets and the second model that is Reason Candidate and Background LDA (RCB-LDA) which extract greatest representative tweets which is obtained from FB-LDA as reason candidates for interpretation of public sentiments.
Keywords- Twitter; Public Sentiments; Sentiment analysis; Event tracking; Latent Dirichlet Allocation (LDA); Foreground and Background LDA; Reason Candidate and Background LDA.