Research Of Drama Hits Ranking Based On The Combination Of Social Network And Search Data

In recent years, Video on demand sites has become the main way for users to chase drama. Premiere on Network and TV at the same day has become a usual way of for dramas. Hits are a major component of the video site’s revenue, and hot series will have higher hits. But the prices of Video’s broadcast copyright are rising, thus, how to predict the drama’s hits in advance is crucial and worth studying.With the development of social networks, more and more dramas were propagandized in Sina Weibo. Bloggers only need to add drama’s public account or related actors, directors as a friend, and then he will get the latest news about the drama and other audiences’ opinions. For the Search engine, it will return the detail information and online playing links about the drama. If we can use these network data to predict the popularity of dramas, then the cost on broadcast rights purchase can be reduced.At first, we analyzed the sentiment orientation of the drama comments in Sina Weibo, and proposed a sentiment orientation analysis algorithm TT-OM which merged the difference on comments object and TV comments pattern. Then we analyzed the drama Hits from the following aspects:Micro-blogging popularity, search popularity, starring and director popularity, comments sentiment orientation, etc. By correlation analysis, from the proposed potential influencing factors of drama Hits, we identified several significantly related factors. And we found important difference between social network data and search engine data in the predicting effects of different time periods. Based on this, we integrated the social and search data, proposed a hits popularity ranking predicting algorithm, the Spearman coefficient between our rank prediction and the real rank is 0.82**and 0.89** in Yoku and iQiyi.