中文摘要：

Web服务的聚类能够改善基于服务的应用如服务发现、组合和QoS预测等．然而目前的聚类方法在相似度度量和信息预处理方面存在一些不足．提出Web服务的QoS和功能两种相似度模型，从不同角度度量服务间的相似度．在此基础上，提出一种特殊的考虑到编程风格和命名规则的预处理方法．最后结合SCAN算法实现了本方法并设计了对比实验对提出的方法进行验证．实验结果表明提出的模型和方法能够有效地提高Web服务的聚类效果．

英文摘要：

Clustering of web services can improve the web service-based applications such as service discovery, composition and QoS prediction. However, the existing clustering approaches have some drawbacks in similarity measuring and information preprocessing. In this paper, two similarity models are presented respectively to measure the QoS similarity and functional similarity between web services. Based on the models, a special preprocessing approach is proposed, which considers the programming style and naming rules. The proposed approach is combined with the SCAN algorithm and evaluated through the planned experiments. The experimen- tal results show that the proposed model and approach can effectively improve clustering of web services.