Chinese Movie Comment Sentiment Analysis Based On Hownet And User Likes

In recent years,With the popularization of network,more and more people like to comment on the movies they have seen on the film platform after watching the movies.The analysis of these movie reviews can provide help for consumer decision-making,business marketing,film production and so on.Therefore,it is of great value to mine the emotional information in film reviews.At present,the emotional analysis of film reviews can be divided into two main categories: the method based on emotional dictionary and the method based on machine learning.However,the method based on emotional dictionary often has the following problems: it is impossible to accurately judge such texts with commendatory or derogatory words;using emotional dictionary and formula to calculate emotional scores,many negative words,degree adverbs and emotional words will be ignored.Supervised learning relies on a large number of labeled data,which makes the system based on supervised learning pay a high labeling cost.Although unsupervised learning costs a lot,the classification results are not very good because of the complexity of Chinese texts.Therefore,aiming at the above problems and combining the advantages of the two,this paper proposes a method combining Hownet Emotional Dictionary and user's praise to analyze the emotions of Douban Movie Comments.This paper constructs a dictionary of movie field by mutual informationand introduces the influence of user's comments on movie reviews.It uses SVM and Logistic Regression and NB training data and makes an example comparison.The method is validated by introducing a new comment data set,which achieves a better classification effect than the traditional method based on emotional dictionary.