Font Size: a A A

Local Expansion Based Community Detection Algorithm Research

Posted on:2017-02-18Degree:MasterType:Thesis
Country:ChinaCandidate:K YinFull Text:PDF
GTID:2308330485961035Subject:Computer technology
Abstract/Summary:PDF Full Text Request
We now are in a world surrounded by complex networks. Many complex systems in real world can be abstracted to complex networks. Complex network analysis draw more and more attention in both academia and industry. The community structure in complex network reveal the high level relationship of the nodes in network, to study a complex network’s community can make researchers to better understand its structure and function.The problem of community detection has been well-studied for decades, a large amount of algorithms have been proposed. However, most of these algorithms require global structure information of the network, and this requirement is getting difficult with the increasingly large scale of the network. Local community detection methods are proposed to solve this problem witch only use the local structure information. In this paper, we propose two new local community detection algorithm for solving some problems which widely existent in many other local community detection algorithms. The main work of this paper is as follow:1) Analysis of a variety of existing local community detection algorithms, uncover different types of node misallocation problems via the community expansion process. Explain the reason of these problems are the algorithms cannot distinguish the inner edges and external edges in the real community structure. To solve this problem, this paper proposed a new type of edge metric based on edge coefficient. This edge metric is used in both two algorithms proposed in this paper.2) Proposed a new type of local community detection algorithm Core Edge based Local Community Detection Algorithm (CELCD) to solve the node misallocation problem. CELCD use core edges as expansion seeds and a new fitness function is presented.3) The expansion processes of the most local community detection algorithms today are based on single fitness function, this paper studied the disadvantage of single fitness function expansion, and proposed a Multi-Object based Local Community Detection Algorithm (MOLCD).
Keywords/Search Tags:complex network, community detection, local community detection
PDF Full Text Request
Related items