Opinion Group Detection Method Research And Realizing In On-line Debate System

With the rapid development of Internet and computer technology,big internet companies have sprung up in today’s China,such as Baidu,Tencent and Ali.At the same time,people’s life and the way of production is constantly improving.people could do shopping,study and work without leaving room.Nowadays,internet companies are no longer just the connectors of traditional manufacturing enterprises,they are also producers,at the same time,traditional enterprises are changing by adding internet attributes to themselves.Today’s business societies are trying to meet all the needs of people,some companies even find ways to create demand for users and then meet it.The internet has been a great power of China’s economy,and how to make some contribution to the rapid development of internet technology,is worthy issue for every software students.Human live in society,their living and production cannot be separated from communication and cooperation,it is everyone’s wish to find himself like-minded people in the so many people.In facing with such a situation,this paper proposes to construct On-line debate system,this system provides users with social platform,promote users to participate in the debates in various forms and express their point of views.Through requirements analyzing and system designing,On-line debate system realized multiple modules such as authority management,online voting,hot topic retrieval,opinion group detection,friend recommendation and online debate.This paper did a deep research on opinion group detection module.At present,though most of the online communication platform,such as QQ,We Chat,Twitter,social tools such as the Line,could group users,their grouping method mainly based on the scope of people’s social relations as well as the user’s interest.Research on grouping according to point of view is rare,thus,this paper makes a deep research upon opinion group detection module.In order to solve the problem of ”group people according to their point of view”,this paper proposed “Bi-directional belief graph based opinion group detecting method”,and developed an on-line debate system with our method.This system encourage users to express their point of view,then group users according to their point of view,finally,find users those who share same opinions with themselves for friend recommendation.According to requirement analysis and system designing,functions such as log in,voting and hot point exhibition are enabled.During the introduce of “bi-directional belief graph based opinion group detecting method”,this paper gave the mathematic definition of opinion group,then analyzed the characteristics of opinion group detection task by comparing it with other clustering method,finally designed opinion group detection method according to the characteristics.In total,this paper analyzes the characteristics of the online debate,build up online debate model,and then use this model to generate a virtual processed chat data results for testing.The results of test turn out to be satisfying.In addition,in order to expand the verification opinion group detection algorithm,opinion group detection method is applied to the literary masterpiece Hamlet to divide characters into different groups.After manual labeling.this paper establish bi-directional belief graph for each chapter in the book,and according to the bi-directional belief graph,extract opinion groups for each chapter,at last extract opinion groups for the whole book.The results obtained turn out to be consistent as readers’ common understanding.When applied,On-line debate system can automatically group user’s into opinion groups according to the discussing record.In general,On-line debate system introduces the hot topic to the user,provides the online debate platform for the user,carries on the friend recommendation to the user,group users according to debate records,those satisfies users’ requirement of acquiring information,on-line discussion and on-line social communication.