The Emotional Analysis Of The Comment Text Of The Bean Film

In recent years,under the leadership of General Secretary Xi Jinping,China's economy has shown a steady and sustained growth trend.Our country has attached great importance to the cultural industry while focusing on economic development.The scale of the film market is also gradually expanding.It has become a necessity for our modern life.At the same time,Internet technology is developing at a high speed,and social networks have gradually penetrated into various fields of people's social life.Various social forums and websites have sprung up and prospered.People prefer to publish their own affairs,entertainment,and sports on the Internet platform.As for the view of the event,the movie review website represented by Douban,Time Network,etc.has accumulated a large amount of user comment text data.Through the user's comment text data,they can know their emotional tendency,and can further analyze the user's purchase behavior to support their purchasing decisions,and also provide positive or negative feedback to the product provider.The film and television industry is moving in a better direction,maximizing the value of user-reviewed text data,which requires crawling user comment text data through reasonable technical means.In this paper,the sentiment analysis and topic extraction of the Douban movie Top250 comment text data are carried out.The specific work contents are as follows:Firstly,this paper uses python web crawler technology to capture a certain amount of comment text data of Douban website as the research object.On this basis,the text data is cleaned and preprocessed,including invalid comments on the text and deletion of abnormal characters.jobs.Then the preliminary processing of the comment information: word segmentation,constructing an emotional dictionary,stop word removal,manual screening,and the like.Secondly,it conducts a further emotional judgment and analysis on the information that has been initially processed.It mainly extracts the nouns,adjectives,verbs or other part of the elements with emotional orientation characteristics from many commentary texts.I built myself or a sentiment dictionary that my predecessors have studied,and designed a reasonable scoring mechanism to calculate the emotional comprehensive value of each comment,so as to judge the emotional tendency of the comment.After that,the innovative LDA unsupervised machine learning technology was applied to the Douban movie review document collection to identify hidden topic information.Converting the comment text in each movie document into digital information that can be recognized by the computer and easy to model.Each comment text represents a probability distribution of some topics,and each topic is used to represent many words.The resulting probability distribution forms a three-layer structure containing words,topics,and documents.Through such a three-layer Bayesian probability model,the theme of each movie review is selected to reflect the theme of the movie.In recent years,more and more scholars have conducted emotional analysis by excavating Internet text comments.As an important part of people's spiritual life,film reviews have also been the focus of scholars.Internet users can learn the whole movie by emotional analysis.Emotional tendencies,whether for the ordinary mass group or the small-scale pursuit of individualized viewing groups,their viewing needs can be met,and relevant government departments can also effectively monitor and control social media through sentiment analysis.