论文部分内容阅读
提出了一种基于WordNet本体标注和概率潜在语义分析(PLSA,ProbabilisticLatent Semantic Analysis)的语义Web服务发现方法OntoPLSA.首先使用WordNet本体标注Web服务的操作名、参数以及用户请求,以经过标注后的输出参数集合为词汇集,服务描述文档集合为文档集,组成词汇-文档矩阵,以该矩阵为输入,使用PLSA方法对服务集进行分类,并将用户请求带入PLSA模型,确定其所属的类;然后在类中以标注后的输出参数为键,含有这个输出的服务的列表为键值,建立一个映射表,查找与用户请求的输出相似的映射表键,进而找出对应的键值,即服务列表;最后根据QoS(Quality of Service)和用户请求中的输入参数确定满足条件的服务结果集合.在415个Web服务组成的数据集上的测试结果表明,性能较其他方法有优势,召回率和R准确率也得到了改善.
This paper proposes OntoPLSA, a semantic web service discovery method based on WordNet Ontology Labeling and Probabilistic Latent Semantic Analysis (PLSA). Firstly, using WordNet ontology, we mark the operation name, parameters and user requests of web service, The parameter set is a vocabulary set, the service description document set is a document set, a vocabulary-document matrix is formed, the matrix is taken as an input, the service set is classified by using the PLSA method, and the user request is brought into the PLSA model to determine the class to which it belongs; Then in the class to the marked output parameters as a key, the list of services containing this output as a key to create a mapping table to find the output of the user request a similar mapping table key, and then find the corresponding key, that is, Service list.Finally, the set of service results satisfying the condition is determined according to the quality of service (QoS) and the input parameters in the user request.The test results on the data set consisting of 415 Web services show that the performance is superior to other methods, and the recall rate And R accuracy has also been improved.