| With the rapid development of the network,people,the mainstay of information dissemination,are increasingly involved in various community media.In recent years,various community tools are also emerging,such as domestic We Chat,QQ,Weibo,Post Bar,Community,etc.,foreign Instagram,Facebook,Line and so on.People read news,share what they see and hear and interact instantly on these social media.Due to the rapid expansion of information and the emergence of different information,it provides conditions for the emergence of spam.These spam can affect the user experience,even affect the stable development of society.Therefore,the treatment of spam has become one of the hot issues in the industry.Most of the treatment of spam in the literature is concerned on two aspects:one is the detection of spam,the other is the filtering of spam.Although these methods can solve spam largely,they are all measures after spam is generated.This paper uses a variety of community discovery algorithms to analyze users and discover potential spammers through known spammers in order to curb spam from the source.In this paper,we use the aggregation coefficient to analyze the aggregation degree of Weibo data,and then use the LPA algorithm,Fast Unfolding algorithm,Simulated Annealing algorithm and Random Walk algorithm to analyze the microblog user data,and get the four algorithms’community division results.According to the divided communities,take the top 10 communities of spammers.The intersection of possible potential spam users obtained by each algorithm is obtained,and 67 lists of user IDs with higher possibilities are obtained.After the above empirical analysis,it is found that the method of discovering potential spammer built in this paper can obtain some suspected list,which has certain reference significance for the research of spam processing. |
