Font Size: a A A

Research On The Application Of Internet Public Opinion Topics Identification And Emotional Tendency Analysis

Posted on:2019-05-29Degree:MasterType:Thesis
Country:ChinaCandidate:Y YangFull Text:PDF
GTID:2417330548473544Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
With the development of the Internet,the media properties of micro-blog are more and more abundant,and more and more people tend to use micro-blog to expose the bad phenomena in the society and to vent their emotions on the Internet.In micro-blog,the views and emotions of Internet users will affect the development of events.If negative public opinion is not controlled and guided in time,the negative information will spread and affect social harmony.Therefore,the topic identification,sentiment analysis and the exploration of the emotional changes of Internet users will be carried out.The characteristics of Internet are of great theoretical and practical significance in monitoring and guiding Internet public opinion.Based on the study of the method of topic recognition and emotional tendency analysis of network public opinion,this paper analyzed the influence of topic polarity on Netizens' emotion and the changing characteristics of netizens' emotional tendencies.The main work is as follows:(1)this paper tried to apply the improved K-Means clustering method based on LDA.To the micro-blog sub topic recognition,and compared with the traditional K-Means clustering method,the experimental results showed that the improvement of the K-Means clustering method based on LDA can identify the micro-blog sub topic,and the effection is better than the traditional K-Means clustering.(2)on the basis of the existing emotional dictionary,we introduced the construction methods of the 7 dictionaries in detail,and make an emotional scoring strategy.Then we used the random forest classification method and the dictionary based method to carry out the emotional classification experiment.The experimental results showed that the dictionary and the emotional scoring strategy constructed in this paper have one in the emotional classification.Set the superiority.(3)combining the evolution stage of network public opinion,taking "Hainan sky-high price ticket" as an example,this paper analyzed the influence of topic polarity on Netizens' emotion and the changing characteristics of netizens' emotional tendencies.It is found that the polarity of the topic has a certain explanatory function on the changes of netizens' emotional tendencies,and in the germination period,the information of public opinion is small and divided.The emotion of netizens is not significant,but in the period of growth,if there are more negative topics,the diffusion of micro-blog information is mainly transmitted,and the negative emotions of netizens will gather quickly.Nowadays is the critical period for government or enterprises to take measures to control and guide the development of public opinion.In the outbreak period,opinion leaders' views played an important role at the development direction of public opinion.The government or enterprises need to focus on the words and deeds of the micro-blog opinion leaders.After the outbreak,the netizens are more sensitive to the new trends related to public opinion events,and the government or enterprises need to continue to pay attention to the dynamics of public opinion.
Keywords/Search Tags:Internet public opinion, Topic recognition, Emotion dictionary, Stage of public opinion evolution, Emotion classification
PDF Full Text Request
Related items