Emotional Analysis Based On Douban Suspense Film Review

With the development of society,more and more people begin to pay attention to TV programs,especially those that are entertaining.Among them,suspense TV dramas are particularly popular,with many viewers attracted by their thrilling plots and exchanging their thoughts on online platforms.Douban,which started out as a platform for sharing books,videos and music,has a large number of users discussing their experience of watching TV shows on the platform.If we can dig out the emotional tendency from these comment text data and conduct in-depth research on it,we will be able to better extract users’ general feedback on TV dramas,provide guidance and suggestions for the creation of TV dramas in the future,and improve users’ experience of watching TV dramas.Based on text data,this paper studies the emotional tendency of suspense drama criticism.First of all,this paper introduces the current research status of sentiment analysis and elaborates the methods of topic analysis,text vectorization model and machine learning.Secondly,Scrapy framework has been used to collect user review data of suspense films on Douban,collecting more than 10,000 pieces of data,and eliminating and reprocessing the data to obtain “clean”data for analysis.By visualization method,descriptive statistical analysis is carried out on fields such as user publication time and user rating.Through the text word frequency statistics,the most popular words about this play were obtained.Through the LDA theme analysis,the subject keywords of positive and negative comments are obtained,so as to understand the reasons for different types of comments given by the audience.Finally,text sentiment analysis is carried out based on machine learning and deep learning methods.After converting user comment text data into data that can be recognized by the model through the bag of words model,naive Bayes and support vector machine methods are used to model the training set data,and the accuracy rate,accuracy rate,recall rate and F1 score are used to evaluate the excellence of the model.The deep learning method based on keras framework is used for emotion analysis,and the appropriate activation function and hyperparameter are selected within the framework.By comparing and analyzing the different evaluation indexes of the fitted model,the appropriate model is selected,and then the emotion classification of the new text data is analyzed with it,so as to provide references for the production and broadcast of domestic suspense dramas.Through analysis,this paper draws the conclusion that the audience focuses on the plot of the story,the rhythm of the film,the ending of the film and the acting of the male and female protagonists.Based on the above analysis,this paper puts forward the following suggestions for the production of suspense film: novel topic selection or reflect real events;The actor’s characterization is full,and the “pit”set by himself in the early stage should be filled completely in the later stage,or an open ending can be set to leave part of the audience’s imagination.