Predicting Success of a Movie from Youtube Trailer Comments using Sentiment Analysis

摘要

The optimal success of a movie can be predicated using social media analytics. The audience expectations and movie buzz can be measured from the data gathered via social media platforms. Few researchers have tried to demonstrate by analyzing and scrutinizing various features of tweets sent during the movie release, that twitter-based prediction of box office revenue are more accurate and specific than market based predictions. Video-sharing websites like YouTube encourages interaction among its users through the provision of user reviewing facility of movie trailers. In this paper, we have categorized the comments of movie trailers on YouTube. Using these comments, we have defined three different Sentiment Indexes which measure the sentiment of movie reviews. Prediction of box office collection of movie is then done based on this sentiment index. This helps the distributors and movie-makers to ascertain the response rate for the movie in advance by observing the comments on the trailers. This also leads to prediction of the release-day earnings by considering at the present-day sentiments. In this paper novel approach of sentiment analysis and sentiment index are used to predict movie collection.