A Trust Model For Wikipedia

As Wiki-based mass collaborative creation and open collaborative communities have been popularized in the new generation of Web,open content is not only widely welcomed in social media,but also attracts attention of the academic communities.Wikipedia is the most famous Wiki site,multi-language encyclopedia based on Wiki technology.Obviously,compared to traditional organization system and other web2.0 applications,Wikipedia provides users with more freedom.But Wikipedia do not focus on individual contribution and lack of obvious mechanism of credibility assessment and trust management.Recently,because of the controversy between the author and the existence of troublemakers,The entry of Wikipedia have been questioned.Variety of users and quick update of content also make the problem of credibility in open collaboration fairly complicated.Based on the relevant research,this paper carry out an expanded innovation research on the trust model for Wikipedia.The main work includes the following two aspects:(1)A structure and edit history based trust model for Wikipedia.On the basis of an edit history based trust model,and the latest research literature on the effect the credibility of open content,this paper builds a trust evaluation model for editors and entry which considers the number of reference and image in the current version of the article and the editor's effective text size.Trust model for entry can able to understand the process of entry change in micro level,and the trust value calculated dynamically.To the editor trust model,it not only can encourage high trust editor warn that low trust,but also to punish the troublemaker.In addition,it can assist the administrator automatically discover high or low trust of the author to decide to increase or decrease the author permission.(2)A two-mode social network based trust model for Wikipedia editors.This paper expresses the relationship between the Wikipedia editors and entries as a two-mode social network and builds a new trust evaluation model for editors through the comprehensive analysis of this network.According to the different class,the algorithm calculates the author trust value considering the author's relative amount of editing,the editing error factor and the clusters coefficient of the entry network under each class.It shows that the model is able to distinguish editors trust under different classes and high trust editors or low trust editors in one classes.