Research On The Correlation Between User Reviews And Movie Box-office Revenue Based On Sentiment Analysis

With the sustained and rapid development of the national economy,the development environment of China’s film industry has been continuously optimized and improved.The related research on movie box office has also received a lot of attention.From the perspective of sentiment analysis,this thesis studies the correlation between the emotional tendencies of movie user reviews and movie box-office.The research of this thesis mainly focuses on two aspects.One is to compare and analyze the emotional indicators of movie users’ emotional tendencies,and the other is to study the relationship between the movie box-office and the emotional tendencies of movie user reviews.The research results show that the user ratings are consistent with the sentiment probability value of the Douban short comment text extracted by the sentiment analysis function of the Snow NLP library,and they can both be used as user tendency indicators.Moreover,the user rating is more accurate in judging the user’s true emotional tendency,and can better reflect the user’s true emotional tendency.The research further introduces multiple linear regression and random forest regression,and deeply explores the correlation between user reviews and movie box-office.At the end of the thesis,according to the research results,the paper puts forward reasonable opinions and suggestions for the research of the movie box office.