Research On Community Detection Method Of Node Structure-Attribute Fusion

There are various networks in the real world,and each network has its own meaning and internal connection.Whether it is the network of relationships between people in life,biological networks,protein networks or academic networks,it can be the object of network research.These networks have high correlation and overlap and are worth exploring and researching.With the rapid development of computer technology and the increasing popularity of the Internet,various social networks have begun to emerge in large numbers,and the analysis of these networks has become more urgent.community detection is a means of network analysis.community detection algorithms facilitate the discovery of communities and the mining of network graph data.Many existing community detection methods use an adjacency matrix to represent the network structure.However,these network structures cannot specifically show the structural closeness between nodes.Second,existing community detection algorithms take less into account the attributes of the nodes and the meaning within the community.In order to solve these two problems,this paper proposes a new algorithm from the perspective of the tightness between nodes and the attributes of nodes.This paper elaborates on the meaning of community detection,and then gives relevant definitions and knowledge of community detection.It will show the current research status of A at home and abroad.On this basis,each community detection algorithm is elaborated,and the advantages and disadvantages of each algorithm are analyzed.Finally,according to the existing algorithms,the proposed algorithm is proposed.Aiming at the problem that the community finds that the structural tightness between nodes can not be reflected when using the neighbor matrix,this paper proposes a R-hop-based node structure tightness matrix calculation algorithm.The algorithm solves the problem that the adjacency matrix can only show the neighboring conditions of the nodes.The R-hop-based node tightness matrix calculation algorithm can show the structural tightness between nodes in addition to showing the connection state of nodes.The community detection algorithm takes less account of the attributes of the nodes and the meaning within the community.Since the object represented by the node has attributes,when calculating the tightness between nodes,this paper not only calculates the structural closeness between nodes,but also calculates the attribute tightness between nodes.In terms of node attributes,in order to measure the importance of different attributes,this paper calculates the weight of each attribute according to the diversity of each attribute value.In this paper,we give a method to calculate the attribute weight and calculate the attribute tightness between nodes.After getting the attribute tightness and structure tightness between nodes,this paper combines the two tightness by assigning different weights to the two tightness and gets the tightness between the two nodes.Then,this paper looks for community centers and initial communities and marks them.Finally,a community classification formula is derived by semi supervised method for community detection.