基于模体的汽车行业站点网络复杂性测度

为了研究大规模网络结构复杂性测度方法,并针对汽车行业站点网络布局与结构功能优化提出对策,基于万维网页面链接数据,构建汽车行业站点网络拓扑结构图.借助VOSviewer聚类算法及Gephi检测并划分网络社团结构,解析基于主题搜索的汽车行业站点内容分类体系及功能结构,并利用Pajek验证各内容社团结构的小世界性,基于Rand-ESU算法检测各社团的模体结构,提出基于模体的网络结构熵算法测度各社团的复杂性.最后,得出汽车行业站点网络社区中模体结构具有同构性,导致社区结构的信息传播功能具有相似性,模体规模与模体信息传播途径多样化对网络结构复杂性影响的显著性较高.

Complexity Measure of Automobile Industry Site Network Based on Motif

In order to provide the methods for measuring the complexity of huge scale networks and put forward the countermeasures for structural layout and optimization of automobile industry site network, the automobile industry site network topology was built based on the linked data of WWW pages. The clustering algorithm of VOSviewer and software Gephi were used for community detection, and the result demonstrates the content classification system, the functional structure and the small-world property of automobile industry site networks. The Rand-ESU algorithm was used for motif detection, and the networks structure entropy algorithm based on motif was proposed to measure the complexity of each communities. The results of empirical analysis show that the network motifs structure is isomorphic among the different communities of automobile industry site network, which have led to the function similarity of information dissemination in different communities, and the motifs size, the diversified channels for information transmission have an important impact on complexity of networks' structure.