Mining online reviews and tweets for predicting sales performance and success of movies

Abstract

Now a day's, online reviews has become one of the most important part of any business. Posting reviews online for products bought or services received has become a trendy approach for people to express opinions and sentiments, which is essential for business intelligence, vendors and other interested parties. Social media like Twitter contains rich information about people's preferences. In this paper movie domain is considered for data mining. For dealing the problem of review mining for predicting sales performance, we have used massive collection of online reviews from IMDB along with thoughts of people about movies from Twitter; both are mined and processed by applying various algorithms. Our analysis shows that both Sentiments captured from reviews and tweets have a major impact on the future sales performance of the movie. For the sentiment factor, we have used Sentiment Probabilistic Latent Semantic Analysis (S-PLSA) model it is a probabilistic approach to analyzing the sentiments in reviews and tweets, which provides succinct summary of the sentiment information. The summarization of both reviews and tweets in terms of sentiments along with additional input of the box office collection from IMDB is given to an Autoregressive Sentiment-Aware model (ARSA) for sales prediction. After that we have done sentiment analysis of online reviews. In sentiment analysis reviews are classified into negative and positive. We used a simple metric P-N ratio to predict the success of movies i.e. Hit, Average, Flop.