Identification And Research Of Abnormal Users In Social Network

In recent years,the identification of abnormal users is one of the most important research fields in the social security field.In order to identify abnormal users in a social network,this paper first proposes the abnormal user identification method based on the basic characteristics of the user,the method uses G-N community discovery algorithm divides the users into two types of community and isolated points,combines rough set theory to calculate feature weight of the user,calculatesmeaningful trust value of characteristicbased on the weights of features,combines with the weights of features and trust value of characteristics calculate credibility of user,the credibility can be used to identify abnormal users.The experimental results of the method show that the better the performance of recognition and the higher recognition efficiency with increasing data set.Secondly,the abnormal user identification method based on the trust relationship between users,the method calcutes the degree of trust relationship based on the characteristics of trust relationship between users,introduces the attenuation factor to reduce the attenuation factord in trust propagation,calculates the user reputation with the trust chain of transmission,judges user class based on the user's reputation.The experimental results show that this methodhas good recognition performance in both case: fewer data sets have been markes and abnormal user data is not balanced in the proportion,compared with the traditional classification method it presents strong stability,accuracy and high efficiency.