Mashup服务网络及其拓扑结构分析

基于从Programmable Web所爬取的原始数据(包括Mashup服务、Web API,以及Tag信息),采用Jacard相似度系数计算2个Mashup服务间的相似度,构建出了大规模的Mashup服务网络,并引入复杂网络理论中的度与度分布、平均最短路径长度、聚集系数以及网络密度与集中性等指标对Mashup服务网络进行度量与分析.结果表明:该Mashup服务网络是一种节点度差异较大、平均最短路径长度较小、聚集系数较高的复杂网络,具有"小世界"和"无尺度"的特性,为后续的基于Mashup服务网络的服务分类与发现、推荐与组合有很好的导向作用.

Mashup services network and its topology analysis

Based on the crawling,the data was rawed,that inclading Mashup service,Web API and Tag information from ProgrammableWeb Site,the similarity between any two Mashup services was calculated by Jacard similarity coefficient and so a large-scale Mashup services network was built.Meanwhile,the index of degree and degree distribution,average shortest path length,clustering coefficient,network density and centralization were measured and analyzed to the Mashup services network.The results show that the Mashup services network is a complex network that node degree is lager,average shortest path length is smaller,clustering coefficient is higher,and have the feature of small-world and scale-free,which provide a good guidance to service classification and discovery,recommendation and composition.