Research On Personalized Mobile Spam Filtering

With the convergence of mobile networks,smart phone terminals and social networks,mobile social networks are becoming more and more popular,facilitating information exchange and information sharing among mobile users,and also leading to the spread of advertisements,rumors and false information in mobile social networks,the filtering of network information can delay or prevent the spread of spam.As most of the previous mobile spam filtering technologies were considered from the spammer's sender or spam content,the individual characteristics and group characteristics of different users were not considered,resulting in the inability of the user's personalized information and social relationships to be fully utilized.As a result,spam filtering results for individual users are not accurate.This article first outlines the concept of mobile spam,the related content of mobile social networking,and the research status at home and abroad.The technologies used in this project,such as social computing,user similarity,crowdsourcing,and group intelligence perception,were performed.Detailed introduction,and summarizes the research progress of the traditional spam filtering technology.In order to solve the problem that the traditional spam filtering technology can not fully utilize the user's personalized features,this paper divides the mobile spam information into two categories: common mobile spam and personalized mobile spam,and then proposes a personalized mobile spam filtering mechanism.For two types of different spam information,Bayesian filters and personalized spam filters based on user interests are used for filtering,and the two are combined to complete the filtering of mobile spam information.Bayesian filters are used to filter common mobile phone spam.The personalized spam filter uses similarity to quantify the degree of similarity between users.The spam information is shared by users and their friends.The spam reports from different users constitute local users.Personalized filtering of mobile spam.In order to solve the problem of similarity calculation,this paper analyzes common similarity calculation methods.In order to adapt to the application scenario of this paper,a similarity calculation method based on information entropy is proposed to calculate the similarity between users.For the calculation of similarity and the need to push spam reports,the system stores user interest information and spam reports in the form of data tables stored locally by users.Finally,in the simulation experiment,this article uses the WeChat user relationship data set to build a virtual mobile social network,and applies a personalized mobile spam filtering system to the network,analyzing the effect of the filtering mechanism and the impact of system parameters.The experimental results show that the system has higher filtering accuracy than other methods and has better resistance to poison attacks.