A Study of Sentiment Analysis Task and It's Challenges

Reviews act as a valuable source of information for decision making. Online e-commerce sites has provided their users to make their opinion about products and services. Huge amount of such opinions are publicly available in the form of reviews. Manufacturers, retailer as well as customers have great interest in customer reviews. A customer has interest in such reviews to determine which product or a particular brand to buy. Manufacturers can analyze the improvement area in their product from such opinionated reviews. Due to large number of reviews available on internet for analysis, it is not cost worthy to read these manually. To optimize this time consuming task there is a need of an automated system which provides summarized result of user sentiments. Sentiment analysis (SA) is the field of study that analyzes people's sentiments or opinion from reviews or opinionated text. In sentiment analysis process machine learning is used to analyze sentiments, emotions and produce a summarized result for decision making. Sentiment analysis can be viewed as a natural language processing task, the task is to develop a system that understands the people's language. Sentiment analysis is a difficult task due to ambiguous nature of human languages( like English). In this paper we present a comprehensive study of sentiment analysis tasks and the challenges for sentiment analysis system. As the aspect based sentiment analysis is an active research area in SA, we will cover this level of sentiment analysis in detail.