Research On The Improvem-ent Of User Portrait Alg-orithm Based On Complex Network

In the past decade,the increasing popularity of increasingly rich mobile applications and Web3.0 applications is prompting users to use mobile devices.Ubiquitous mobile devices record digital traces of different human activities,such as movement,purchase transactions,preferences,opinions,etc.They are all important sources of information for studying environmental monitoring,transportation,social networks,innovative demographic indicators,and human behavior.As a result,social network research has undergone tremendous changes,and the study of mathematically complex social network information,which is famous for graph theory,has attracted the attention of many disciplines.We have seen the emergence of social network science,which is an attempt to understand the various networks that emerge in nature,technology,and society with a unified set of tools and principles.The user portrait system has risen rapidly with the explosive growth of data.In different fields,user portraits are often used to serve a variety of basic tasks,such as precision marketing,data statistics,deep mining,product design,user behavior&industry reports,etc.Today’s user portrait algorithms mostly attribute the user’s personal information or the user’s text information to the user’s attribute labeling classification process,and provide the same service to users with the same attribute label.However,it lacks a labeling algorithm based on user behavior and user’s position in the process of information dissemination.Based on the k-shell algorithm,Weaver-Thomas model and other methods,this thesis proposes the k-Gr algorithm applied to the direction of user portraits.The direction of the algorithm is to perform layered processing on large and complex networks based on user behaviors in order to find key nodes at the source of information dissemination in the network.Through the k-Gr algorithm,different users are divided into ordinary users,intermediary users,star users and authoritative users.In the end,this article also puts forward corresponding suggestions for the customer operation of the four different types of users in the process of information dissemination,in order to enable the enterprise’s information to be effectively and quickly spread on the network.