Evaluating Performance of Machine Leaming Techniques used in Opinion Mining

Sentiment analysis is a computational study used for interaction of humans with the computers. In the task of Opinion Mining computations are done on words and sentences used by individual for expression of their opinions. This computation is done with the help of NLP tools to determine whether the review or opinion expressed is positive, negative or neutral. Binary categorization (Positive, Negative) and Multi class categorization (Positive, Negative and Neutral) are two main text classifications used for analysis. Nowadays Opinion is not limited to an individual, it is also an important aspect for most of the organizations as it helps in intensive response from their customers regarding their product quality or about any services used. Earlier, organizations used to conduct different type of Surveys, Questionnaires in order to improve their quality or services offered. There are two widely used approach used for study of opinion mining which are dictionary based approach and machine learning approach In this paper we provided information about different machine learning approach used for task of Opinion Mining, performance overview of various approaches used, metrics used to measure opinion mining efficiency.