Research And Implementation Of User Influence Evaluation System For Social Networks

With the rapid development of the Internet,online social networks have become popular communication platforms.Social network users can share and disseminate various information.Massive users and published information in social networks have great commercial and research value,attracting a large number of scholars to study it.User influence evaluation is one of the important research directions in social network research.User influence evaluation can find influential user groups in social networks,which is of great value for advertising,user recommendations and online marketing.This thesis first proposes an influence evaluation algorithm suitable for dynamic networks for the influence evaluation problem,and then designs and implements a social network user influence evaluation system based on this algorithm.The dynamic network influence evaluation algorithm takes into account the time and dynamic attributes of social networks.Real social networks have a changing process,and user influence will change accordingly.The existing static network influence evaluation algorithms cannot reflect the changing process of social networks.In this thesis,dynamic networks are divided according to the time dimension.When the network changes between time slices,an incremental network model is used to update the influence.The incremental influence evaluation algorithm maintains most of the network structure that has not changed,only calculates the incremental part of the change,reduces a large number of repeated calculations,and improves the efficiency of imfluence evaluation.The dynamic network influence evaluation algorithm in this thesis includes user relationship edge weight algorithm,influence evaluation algorithm,shortest path update algorithm and incremental dynamic network partition algorithm.The user influence evaluation system is mainly based on a data mining algorithm that takes the dynamic network influence evaluation algorithm as the core.The system improves services for advertising agencies and weibo users.Through the system,advertising agencies can efficiently find product spokespersons,and weibo users can analyze the operation of their accounts.The system is designed and implemented according to the division of modules,which mainly includes front-end modules,data acquisition modules,data processing modules,data mining modules,database read-write modules,and log management modules.Finally,this thesis verifies that the social network user influence evaluation system is consistent with the design expectations through system functional tests and non-functional tests.