基于使用信息的Web服务质量评价方法

发布时间：2018-08-06 21:39

【摘要】：随着互联网的迅猛发展,Web服务提供商开发各个领域相关的Web服务,因此会存在大量功能相似的Web服务供Web服务销费者选择。消费者一是要关注Web服务是否满足业务逻辑,二是要关注QoS,即Web服务的质量。而如何在高质量的Web服务候选集中选择出满足用户个性化需求的服务是研究者们广泛研究的课题之一。通过对现有的Web服务评价方法的深入研究分析,针对现有方法只根据Web服务的QoS参数进行评价方法存在的问题,本论文考虑了用户个性化需求因素,提出了一种面向个性化需求的Web服务评价方法WSQEMBUI,从而进一步提高Web服务质量评价的准确率。在该方法中首先从Web服务中介服务器的使用日志中获取各个服务的使用情况数据,然后对数据进行预处理,根据Web服务使用信息数据为每一个服务建立评价概念树,评价概念树每个结点表示服务的一个评价因子,根据已经构建好的服务质量评价概念树找到一条与个性化需求因素接近的路径,即服务候选集,再计算路径上每一个评价因子的权重,最终给出每个服务的评价结果。本文又提出了结合了专家的先验知识的服务质量评价概念树的优化算法,这一优化算法主要是结合了专家设定评价因子权重与机器学习方法。最后,理论分析了本文的服务质量评价方法的可行性。最终通过实验验证了面向个性化需求的Web服务评价算法在一定程度上提高了Web服务质量评价的准确率。
[Abstract]:With the rapid development of the Internet, web service providers develop related Web services in various fields, so there will be a large number of similar Web services for Web service sellers to choose. First, consumers should pay attention to whether Web services satisfy business logic, and second, they should pay attention to QoS, that is, the quality of Web services. However, how to select a service to meet the individual needs of users in the high quality Web service candidate set is one of the topics widely studied by researchers. Based on the in-depth study and analysis of the existing Web service evaluation methods, this paper takes into account the user's personalized requirements for the existing methods, which are only based on the QoS parameters of the Web services. A new Web service evaluation method, WSQEMBUI, is proposed to improve the accuracy of Web service quality evaluation. In the method, the usage data of each service is obtained from the usage log of the Web service mediation server, and then the data is preprocessed, and an evaluation concept tree is established for each service according to the Web service usage information data. Each node of the evaluation concept tree represents an evaluation factor of the service. According to the established concept tree of service quality evaluation, we find a path that is close to the personalized demand factor, that is, the service candidate set. Then calculate the weight of each evaluation factor on the path, and finally give the evaluation results of each service. This paper also proposes an optimization algorithm for quality of service evaluation concept tree which combines the prior knowledge of experts. This optimization algorithm mainly combines the weight of evaluation factors set by experts and the method of machine learning. Finally, the feasibility of the service quality evaluation method is analyzed theoretically. Finally, the experimental results show that the Web service evaluation algorithm for personalized requirements improves the accuracy of Web service quality evaluation to a certain extent.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】