基于语义的Web服务匹配模型的研究与实现

发布时间：2018-06-16 22:17

本文选题：语义 + 本体　；参考：《电子科技大学》2013年硕士论文

【摘要】：在互联网技术及其应用热潮的推动下，Web服务被视为一种全新的Web应用模式而出现，它为异构平台中的资源整合以及资源共享等互操作提供了标准。随着Web服务数量的不断增长，如何帮助用户在候选Web服务集合中找到满足其个性化需求的服务成为一个世界范围内被广泛重视的问题。由于语义级别的服务匹配相比于传统的基于关键字的语法级别的匹配在查全率和查准率上得到了改善，因此语义级别的服务匹配得到了业界人士的认可。但是目前的基于语义的Web服务匹配还不够成熟，存在着一些片面性，诸如：缺乏对上下文信息的考虑、缺乏对服务属性的不确定性的考虑以及在匹配度计算过程中所采用的贪婪算法引起的误报和漏报问题等。本论文针对上述问题，对语义Web服务匹配模型进行了研究，重点研究了如何对Web服务和服务请求进行描述以及如何对服务请求和候选服务进行匹配，继而将最贴合用户需求的服务推荐给用户。本论文在现有语义Web服务匹配方法的基础上，提出了一种混合的服务匹配算法。该算法主要结合了粗糙集的属性约简算法和偶图最优匹配算法的优点，首先，，分别从服务请求和服务广告的描述文件中提取出待匹配信息，并且运用粗糙集属性约简算法对服务属性中的非决定性属性进行删减，然后根据删减过后的服务属性建立偶图，从而将服务请求和服务广告之间的匹配度计算问题转换为偶图的最优匹配求解问题。该服务匹配方法综合考虑了输入、输出、前提、效果等功能信息和上下文信息、用户偏好、QoS等非功能信息，提高了匹配结果的准确率和召回率。最后，本论文基于该混合的服务匹配算法设计并且实现了一个Web服务匹配模型，并且通过实验对该语义Web服务匹配模型进行验证，通过与其他算法的实验结果进行对比分析，说明了该混合的服务匹配方法提高了匹配结果的召回率，同时也说明了上下文信息的匹配可以提高匹配结果的准确率。
[Abstract]:Driven by the upsurge of Internet technology and its applications, Web services are regarded as a new Web application mode, which provides a standard for resource integration and resource sharing in heterogeneous platforms. With the increasing of the number of Web services, how to help users find a service to meet their personalized needs in the set of candidate Web services has become a problem that has been paid more and more attention in the world. Because the semantic level service matching is better than the traditional keyword based syntax level matching in recall and precision ratio, the semantic level service matching has been recognized by the industry. However, the current semantic based Web service matching is not mature enough, and there are some one-sidedness, such as the lack of consideration of context information, There is a lack of consideration of the uncertainty of service attributes and the problem of false positives and omissions caused by greedy algorithms used in the calculation of matching degree. In this paper, the semantic Web service matching model is studied, focusing on how to describe the Web service and service request and how to match the service request and candidate service. In turn, the most suitable user needs the service recommended to the user. Based on the existing semantic Web service matching methods, a hybrid service matching algorithm is proposed in this paper. The algorithm mainly combines the advantages of the attribute reduction algorithm of rough set and the optimal matching algorithm of bigraph. Firstly, the information to be matched is extracted from the service request and service advertisement description file, respectively. The indeterminate attributes in the service attributes are deleted by rough set attribute reduction algorithm, and then the bipartite graph is established according to the service attributes after the service attributes are deleted. The problem of the matching degree between the service request and the service advertisement is transformed into the optimal matching problem of the bigraph. The service matching method takes into account the input, output, premise, effect and other functional information and context information, and users prefer non-functional information such as QoS, which improves the accuracy and recall rate of the matching results. Finally, this paper designs and implements a Web service matching model based on the hybrid service matching algorithm, and verifies the semantic Web service matching model through experiments, and compares with the experimental results of other algorithms. It is shown that the hybrid service matching method can improve the recall rate of the matching results and that the matching of the context information can improve the accuracy of the matching results.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP393.09;TP391.1

【相似文献】