Sentiment Analysis and View Count Forcasting Using Youtube

Abstract

Youtube has become one of the most successful video sharing website which allows the users to view, share and comment on the video posted. In this paper, we are interested in how the emotion expressed in the viewers’ comment would affect the view count of the video. The comment of the selected music video are collected using the Youtube Data api and are preprocessed by removing useless features. The processed comment are then classified into six categories: anger, disgust, fear, joy, sadness and surprise during the sentiment analysis. Generally, we can find certain correlation between the emotion expressed in the comment as well as the view count of the video.