An Approach to Sentiment Analysis on Gujarati Comments on Gujarati YouTube Video

Abstract

For the last several years, YouTube has being the sensation for social videos sharing, also high amount of people post their reviews on shared video. But manual analysis of review classification is very difficult to do and also became time consuming. Also sentiment analysis for English language has expanded very high but for Indian regional languages, the process is very slow. So this paper proposed a technical way for review classification on Gujarati comments posted by users on Gujarati videos. A general process that can be implemented in future for review classification on Gujarati review comment is brought with detailed description. This process will give positive, negative and neutral analysis for video review, so users can easily identify no. of positive, negative and neutral reviews on particular Guajarati video. Data used for Gujarati video review classification has been collected from Youtube.com. As a pre-processing technique tokenization, stop-word removal and stemming and for Feature Extraction POS (Part of Speech) tagging has been used. In this paper for Gujarati video feature analysis hybrid technique has proposed and graphical representation of feature analysis has shown by column chart.