Research And Application Of Conversation Topic Mining And Backtracking Method Based On Interactive Text

The Internet era has spawned many instant messaging platforms(wechat,QQ,microblog,etc.).On these platforms,a large amount of text information is generated every day.The text containing more than two information subjects is called interactive text.From these texts,we can dig out a lot of valuable information,such as social public opinion,business opportunities,potential customers,telecommunications fraud and so on.However,the above information is hidden in a large amount of text.It is necessary to study and apply appropriate intelligent technologies such as data mining algorithms in order to automatically mine the concerned information and apply it to the fields of public opinion analysis,personalized recommendation,public security and so on.Therefore,this paper analyzes the characteristics of interactive text in detail,and puts forward the topic feature model of interactive text conversation;Secondly,in order to realize topic backtracking,a topic backtracking model is proposed;Thirdly,in order to improve the accuracy of topic backtracking,the mainstream feature selection algorithm is studied and optimized;Finally,a real-time information interaction behavior analysis system is designed and implemented.The main work of this paper is as follows;(1)This paper studies and analyzes the characteristics of interactive text,document topic and dialogue topic,constructs a conversation topic model,and puts forward a conversation topic feature model for interactive text.(2)According to the demand of topic backtracking,this paper constructs a conversation topic backtracking model.(3)In order to improve the accuracy of topic backtracking,the mainstream TFIDF algorithm is selected and optimized,and an improved TFIDF algorithm LRIDF algorithm is proposed to realize feature selection.The main improvement is to integrate the length of feature words and the weight of associated words based on entries in the dictionary to improve the performance of the algorithm.After empirical analysis,it has achieved good comprehensive performance.(4)The research results of this paper are applied to the real-time information interaction behavior analysis system,the design scheme of the application system is given,the prototype system is programmed,and the usability of the research results of this paper is verified by the actual data test of telecom operators.The main contributions of this paper are as follows: a conversation topic feature model for interactive text is proposed;This paper constructs a conversation topic backtracking model;An improved TFIDF algorithm LRIDF algorithm is proposed;A real-time information interaction behavior analysis system is designed and implemented.