Revisiting Online Video Popularity: A Sentimental Analysis

Abstract

This research deploys sentiment analysis to investigate the influence of comments and construct a video popularity model. We selected eight channels on YouTube to examine the proposed model in terms of 306 videos and 59,103 comments and we collected 200 samples. The results showed high number of views may not result in high positive comments. The influence of own factors of videos such as title of videos, video thumbnail, and content was higher than objective factors such as number of views, like or dislike, and sentiment analysis. Finally, we discovered sentiment indirectly influences the decision of watching behavior.