摘要
在无标度网络中,社区结构是普遍存在的一种网络结构特性,社区结构是网络中间层的描述,是对网络的自然压缩。文中基于这一事实,将社区结构发现方法加入到多层网络划分框架中,提出了基于社区结构的多层网络划分改进策略。该方法首先对无标度网络进行社区发现;然后以发现的社区结构为单位,对原网络进行压缩;之后对压缩后的网络进行初始划分;最后将划分结果还原为对原网络的划分。在进行初始划分时,为获得较好的划分效果,引入了0-1规划方法,并使用K-L算法进行优化。通过对比实验,结果表明把社区结构引入多层网络划分方法中,可以获得更好的划分。
In the scale-free network,the community structure is ubiquitous structural properties of a network,community structure is the description of network middle layer,which is a natural compression for network Based on this fact,the community structure discovery methods are added into multi-layer network framework,propose an improved multi-layer network division strategy based on community structure. This method first carries out the community discovery for scale-free networks,then with the discovered community structure as a unit,conduct the original network compression,later divide the network compressed initially,finally the result will be reverted to the o-riginal network division. During initial division,in order to get a better division results,introduce the 0-1 programming methods and algo-rithms and optimized by the use of K-L. By comparing the experiment,the results show that introduction of community structures into multi-layer network division method,can get a better division.
出处
《计算机技术与发展》
2014年第11期234-237,241,共5页
Computer Technology and Development
基金
国家"863"高技术发展计划项目(2013AA01A212)
关键词
网络划分
无标度网络
社区结构
多层网络划分
network segmentation
scale-free network
community structure
multi-layer network division