Evaluation of Machine Learning Techniques in Sentimental Analysis

Sentiment analysis has vital applications in several areas, together with selling, recommendation, and financial analysis. It is considered as a process to find the polarity of given data. Extracting sentiments, the helpful contents, and semantics from the opinion sources manually becomes a difficult task in presence of millions of reviews. This paper has a detailed analysis of various machine-learning techniques and compared them on the basis of their accuracy, benefits, and limitations of every mechanism. Experimental results show that supervised machine learning techniques achieve the higher accuracy that is beyond that of unsupervised learning techniques.