Research On Sampling Representative Users Of Social Networks Based On Weighted Neighborhood

With the explosive growth of users on social platforms,how to extract a subset that can represent the characteristics of all other users in the whole network is a very challenging problem.In recent years,scholars mainly use topology based sampling and user attribute based sampling to study the representative user sampling problem,these methods only consider the representative user sampling problem from a single aspect,and the extracted representative user subset can not well represent the whole original network.Based on previous studies,we explore the problem of representative user sampling,and propose the representative user sampling algorithm for social network based on the weighted neighborhood.In our algorithm,the importance of a node is measured by combining the degree of the node itself with the weight of the neighbor and the degree of the neighbor node,and the weighted neighborhood is used to describe the topology of the network.The method mainly includes three steps:(1)All users are divided into different attribute groups according to attribute values,and the representation of users in each attribute group is calculated by attributes and weighted neighborhood;(2)The user's attribute group representation is obtained by calculating the sum of user's representation in each attribute group,and the representation of all user's attribute groups is added up,which is evaluated numerically based on the quality function;(3)Heuristic greedy algorithm is used to extract representative users.On four datasets,this thesis proposes the representative user sampling algorithm for social network based on the weighted neighborhood,which is compared with the other six algorithms,the results show that our algorithm is superior to other algorithms in accuracy,recall and F1-Measure on different datasets.