Research on network users archives matching based on maximum weight matching of bipartite graph

A network users matching model based on maximum weight matching of bipartite graph was presented in the paper. In order to avoid the defect of correctly matched users' profiles missing at attribute value exact matching process, relying on schemaless attribute value similarity, first of all the model selected candidate user set from the ones to be identified, and established a bipartite graph with the attribute value of candidate user profiles and that of source user profiles to be matched. The edge weight of the bipartite graph is subject to Dice comparability coefficient of each attribute value. And then obtained the similarity coefficient of comprehensive profiles of the source users and candidate users. If the comprehensive similarity coefficient of candidate user profile is greater than that of similarity threshold of profile, the candidate user profile is matching with that of source user profile through solving the maximum weight matching of bipartite graph. Eventually, through matching calculation of actual database, the results show that the model can carry out correct matching effectively and specifically for network users of different systems. It has conquered model heterogeneous existing among user information attributes. Through comparing with precise matching algorithm based on attribute value, the algorithm presented in the paper has improved 10% and 5% on recalling rate and accuracy rate respectively. © 2016, UK Simulation Society. All rights reserved.