A semantic Web service discovery method is proposed based on service clustering. The semantic similarity of function and input/output parameters is calculated, and the services in a service library can be clustered by a clustering algorithm. The input/output parameters of a cluster are marked by a unified label, then an input/output concept set is obtained and a unit model of service cluster nets is constructed. Finally, a unit matrix of service cluster nets is proposed, and the favorable Web service could be discovered effectively based on the matrix. The formal model of service cluster net units is presented to process the service discovery. The validity and reliability of the proposedmethodareillustratedbyanexperiment,showingthattheperformanceofvalidityandcompletenessisobviouslyenhanced.% 提出了一种基于聚类的语义Web服务发现方法。通过计算Web服务的功能相似度及输入输出参数的语义相似度，利用聚类算法对服务库中的服务进行聚类。统一标注服务簇的输入输出参数，得到输入输出概念集，构造出服务簇网元模型，并提出标识服务簇网元的矩阵模型，从而实现服务的快速发现，找出最符合用户需求的服务类。给出了服务簇的形式化描述及构建算法，并进行服务发现。实验例证了所提出方法的有效性和合理性，以及在查准率和查全率方面的明显提高
展开▼
