A Survey on the Present State-of-the-Art of Sentiment Analysis Methods

Abstract

Of late, social media has started generating unprecedented data which became a goldmine for researchers and business organizations. The proliferation of Online Social Networking (OSN) applications paved way for people to socialize and opine freely. Thus the data on OSN assumed importance as the opinion of people can influence businesses or help them to know sentiment of people. This knowhow can help organizations to have better strategies to promote customer satisfaction with increased Quality of Service (QoS). Sentiment analysis thus became indispensable part of decision making systems of enterprises as it can provide required Business Intelligence (BI). Though there has been considerable research on this area, it is still open to opportunities, possibilities and optimizations. This paper throws light into the present state-of-the-art of sentiment analysis and finds gaps in the research. It tries to cover the landscape of the subject in terms of lexicon-based methods, knowledge-based approaches, machine learning techniques, classification strategies, ontology-based methods, aspect oriented method and semantic approaches found in the literature in the study of sentiment analysis. This paper also includes study on sentiment analysis in social media such as Twitter and Facebook, real time applications of sentiment analysis besides various measured employed as part of sentiment analysis methodologies. It provides useful insights on various aspects of sentiment analysis that leverage Information Retrieval (IR), Data Mining (DM) and expert systems where BI is extracted and interpreted for making well informed business decisions.