Research And Application Of User Profile Technology For Dark Net

In recent years,in order to prevent anonymous users from spreading false information in the network,damaging the reputation of others or inciting panic emotions,and other cyber crimes,the Cyber Security Law has added a real-name management method based on the user's real name.However,at present,the dark net is still in the blind zone of the national network security supervision.A large number of criminals use the dark net to put on the "invisibility cloak" to carry out various illegal acts.Therefore,it is of great significance to mine user information in the "invisible and intangible" dark network and construct the user portrait of the dark network for the supervision of the dark net.At present,most of the researches on user portraits are only for the users of surface network,but in the anonymous,hidden and weakly structured dark net,there are few researches on user portraits,and it is more difficult to construct rich user portraits.In this paper,we will study and apply how to obtain users' data and build user profiles through sparse information for dark net.This paper builds a Tor-based dark web collection system to obtain user information,transaction records,and social speech data left by users in the dark web,and correlates and expands user data through some dark web user IDs.After the data is cleaned,user features are extracted from the dark web user data sets in both Chinese and English.Analyze structured data,build a dark web keyword vocabulary for unstructured data,and combine natural language processing algorithms such as named entity recognition and entity relationship extraction to obtain a three-dimensional user base consisting of basic information,social behavior,and market transactions The attributes constitute the user's basic profile.In order to further mine user attributes and solve the problem of sparse user attributes on the dark net,we tag all kinds of user characteristics through algorithms such as sentiment analysis,position analysis,influence calculation,activity calculation,and merchant sales forecast.Expand the user's basic portrait into a deep portrait.Using the dark net user portraits constructed in this paper,by calculating the similarity of the user portraits,a clustering algorithm is designed to realize the discovery of dark web user virtual groups.In this paper,we build a dark network user profile with 21 user attributes,including 5 basic attributes,8 social attributes and 8 transaction attributes.The analysis of the accuracy of user attributes is better than that of the recent relevant research applied to the dark net data set.