Research on User Comments of Douban Animation Made in China Based on Text Mining Technology
Siyu Sun, Yingjie Gai, Yingying Zhou, Aiting Xu · 2019
With the advent of the "Internet plus" era, the online film criticism has springing up rapidly. The success of "NeZha" has brought the domestic animation film to an unprecedented high tide and brought a lot of information about the film reviews. Based on Chinese natural language processing technology, this paper takes the Douban movie review of domestic animation series as the research object, uses Python to crawl the movie review data, on the basis of the preprocessing of data cleaning, data normalization, Chinese word segmentation and removal of stop words, etc., carries out machine learning based emotional tendency analysis, and the visual analysis of word cloud and the analysis of sentiment orientation based on machine learning, At last, uses LDA theme model to mine movie reviews in depth. The research shows that: the audience pays more attention to the plot setting, character shaping, picture presentation, production ability, line performance and other aspects of domestic animation series films; Douban users tend to be positive attitude of domestic animation, but there is still a large proportion of negative emotion;the praise and popularity of domestic animated films continue to rise, but the uneven quality of domestic animation films still needs to be improved.