基于历史调用日志分析的Web服务间语义相关性发现

摘要

Aiming at discovering semantic associations among Web serviceB as the critical problem of Web service research, the uncertainty of the semantic association between Web services cannot be well represented by current methods, and consequently cannot be used for automatic service discovery and composition effectively. In this paper, the definition of historical logs of service invocations was given at first. Then, by means of statistical computation, the method for computing the weight of a service component was given based on the PageRank algorithm. Further, the method for measuring the semantic associations between services components upon the electrostatic field theory- Our methods can establish basis for service search, discovery and composition. Experimental results show that our method is efficient and practical.%针对Web服务间语义相关性发现这一Web服务领域研究的关键问题,目前的方法不能较好地描述服务间语义相关的不确定性,进而不能有效地支持自动服务发现与组合.以Web服务描述中的操作作为服务组件,首先给出Web服务调用历史日志的定义,接着以统计计算为手段,基于PageRank算法给出服务组件权重的计算方法,进一步基于静电场理论给出Web服务组件间语义相关度的度量方法,为Web服务搜索、发现与组合奠定基础.实验结果表明,该方法具有高效性和实用性.