基于综合本体相似度计算的WEB服务发现算法

摘要

The main process of service discovery is to match lOPE between the service requests and the registered services.This paper presents a new service discovery frame and an improved comprehensive similarity algorithm.The similarity of services is calculated by comprehensively considering the semantic distance, the depth, density, and information sharing in ontology tree concepts.The service discovery process is divided into two parts, the filtering phase and the sorting phase, and the more adaptable similarity computation approaches are used on different parameters of the services respectively.The services input provided by the users are used in the last sorting phase for diminishing the deviation caused by the inaccuracy of information the users provided.After using SME2 to evaluate the environment and using test set to appraise, the recall and precision ratios are improved to a certain extent compared with other algorithms.%服务发现的主要过程就是对服务请求和己注册服务间IOPE进行匹配.提出一种新的服务发现框架,以及改进的综合相似度算法.综合考虑语义距离,本体树中概念的深度、密度,以及信息共享来计算服务的相似度.将服务发现过程分为过滤和排序两阶段,并对服务的各参数分别采用更适合的相似度计算方法.将用户给出的服务输入用于最后的排序阶段,以减少由于用户给出信息不准确带来的偏差.在SME2评测环境以及测试集进行评价后,与其他算法相比,查全率/查准率有一定的改进.