Font Size: a A A

Design And Implementation Of The Opinion Leader Portrait System In Social Network

Posted on:2019-08-08Degree:MasterType:Thesis
Country:ChinaCandidate:C Y XuFull Text:PDF
GTID:2428330566967194Subject:Software engineering
Abstract/Summary:PDF Full Text Request
The virtual nature,diversity,diversity and freedom of social network make the public more and more interested in choosing social media to express their personal status and exchange information.Research on opinion leaders in social networks helps to analyze in depth the way information is disseminated in social networks.It helps guide the development of public opinions in society and helps the government.Make decisions and resolve related emergencies.At the same time,the identification research and application of opinion leaders are of great significance for realizing accurate advertising based on big data and promoting brand marketing.Therefore,opinion leader research in social networks has very important communication significance,social significance and economic value.The traditional social network opinion leader identification mainly selects the opinion leader from the user attribute structure.The content-based opinion leader identification method mostly quantifies the user attribute structure index of a specific topic and does not change substantially.In order to solve the shortcomings of a single recognition method,which takes into account the characteristics of user structure,user behavior and user topic,this paper proposes a method of opinion leader recognition(HLDAD)based on hierarchical clustering and LDA topic model.In this method,the data of the collected social network user is preprocessed at first;then,the user topic keyword is extracted by a hierarchical clustering and LDA topic model,and the distribution features of the user topic are obtained,thus the user theme community is constructed.Finally,based on the calculation of user structure characteristic value,importance index and activity index,the opinion leader based on topic is identified.Experimental results show that the algorithm proposed in this paper can accurately identify the topic opinion leader.Secondly,in order to solve the problem of opinion leader's management and realize the effective application of opinion leaders,give full play to the social value and economic value of opinion leaders.This paper designs and implements the systemof opinion leader figures.The tag leader attributes,behaviors,and topic features are tagged and finally displayed by a visualization tool.
Keywords/Search Tags:Social network, opinion leader, hierarchical clustering, LDA topic model, user portrait
PDF Full Text Request
Related items