基于Mashup和服务语义的可组合服务推荐  

Novel Composable Service Recommendation Based on Mashup and Service Semantics

作　　者：革家象 刘子豪[2] 许斌[1] 

机构地区：[1]清华大学计算机科学与技术系,北京100084 [2]北京信息科技大学信息安全系,北京100192

出　　处：《小型微型计算机系统》2015年第11期2434-2438,共5页Journal of Chinese Computer Systems

基　　金：国家"八六三"高技术研究发展计划项目(2013AA01A607)资助;国家自然科(61170212)资助;北京多媒体网络化重点实验室资助

摘　　要：互联网上每天有大量的Web服务和Mashup被创建.因此,如何向用户推荐可能进行组合的服务(可组合服务)是一个具有挑战性的问题.现有的一些服务推荐方法存在二八定律,即根据服务组合的历史信息来进行服务推荐的方法会导致流行的服务不断的被推荐,所推荐的服务集中在20%的范围左右.提出一种服务推荐方法,基于已有的服务组合(Mashup)信息和服务之间的语义关联,进行可组合服务推荐.我们首先介绍服务组合推荐的过程和基本架构,然后描述如何预测潜在Mashup里的可组合服务和计算服务之间的语义关联度.通过在包含11182个API和7568个Mashup的数据集上进行的实验,证明了我们的方法有效地扩大了43%的服务推荐范围.A large number of newWeb service and Mashup have been created on the Internet every day. Thus,recommending services which can be composed together（ Composable Service） is interesting but remains a tough problem. Current service recommendation approaches hold Pareto principle,namely recommendation rely on historic services composition information which results in popular services being recommended frequently and the recommended services are concentrated in around 20% of scale. In this paper,we proposed a novel composable recommendation based on two dimensions. These two dimensions involved current composition information and services semantic relatedness calculation. First,we presented service composition recommendation procedures and structure. Then,we describe composable service prediction in Mashup and services semantic relatedness calculation approaches. After experiments on dataset with 11182 APIs and 7568 Mashups,the result shows our approaches effectively extend 43% of recommendation scale.

关 键 词：可组合服务 服务推荐 服务组合历史 语义关联度 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]