A Survey on Sentiment Analysis on Social Network Data

Abstract

Sentiment analysis is an area of research in educational as well as commercial field. The word sentiment denotes the moods or attitude of the person to some particular domain. Therefore it is also known as opinion mining. Opinions of a person may differ from another person. Opinion mining also leads to the particular impersonations on the domain, not facts since the sentiment analysis are mostly topic based. Sentiment classification involves the classification of the polarity and the emotions . Sentiments can be analyzed and classified either by machine learning techniques or by lexicon based techniques. Sentiment analysis allows an user to get a clear idea regarding the “customer satisfaction and dissatisfaction” which For example, “public opinion on new launch of google’s phone” etc. In the commercial world, consumer’s feelings or opinion towards some product or product are very significant for its sell. Therefore in decision making and in real world applications ,sentiment analysis plays a major role. Twitter is considered to be the one of the most populous social networking site where millions of users share their suggestions and opinion about the several fields like politics, products, personalities etc. Many study works are done in the arena of sentiment analysis. But then they are only beneficial in modeling and tracing public opinions. Since the exact reasons behind the sentiment variations are not known and Therefore such variations are not useful in decision making. Sentiment analysis has several applications in various fields like political domain, sociology and real time event detection like Tsunami. Earlier studies were done to model and track public opinions. But then with the advancement in technology, today we can use it for interpreting the reasons of the sentiment change in public attitude, mining and summarizing products reviews, to solve the polarity shift problem by performing dual sentiment analysis. Here we use different algorithms/models like Naïve Bayes (NB) classifier, Support Vector Machine (SVM) algorithm and so on.