Comprehensive Rating Model of Douban Movie Based on Sentiment Analysis
Yiran Gu, Yu-Yu Su · 2017
the rating mechanism of Douban movie simply calculates mean value of all users rating which does not take into account the emotional inconsistency between users rating and their comments. The mechanism fails to differentiate between the impacts on total scores and the movie reviews of different quality. In order to revise the rating, this paper extracts the features of the movie reviews by improved IG algorithm, quantifying the emotion value automatically by Naive Bayes classifier. Based on the above, the paper also constructs a movie comprehensive rating model combined by the edit time of each comment and number of like etc. The experimental results show that the proposed model reflects more realistically and effectively in users' emotional tendency.