Description

Sentiment analysis and opinion mining is an emerging area of research for analyzing web data and capturing the sentiment of the users. This research presents sentiments analysis on Indian movie review corpus using machine learning classifier. Bayesian Classifier has been used in this study for testing feature selection mechanics. This classifier is trained on the words/features of the corpus extracted using five feature selection algorithms (Chi-square, Info-gain, Gain-Ratio, One-R and relief attribute) and a comparative study have been performed amongst them. The classifier and feature selection approaches were evaluated by two different metrics (F-Value, and False Positive). Results of this study show that: for maximum number of features, Relief -F feature selection approach is found to be good with better F-value, Low FP Rate. In addition for the less number of features, One-R was better than Relief-F. This electronic document is a “live” template and already defines the components of your paper [title, text, heads, etc.] in its style sheet.