The Research Of Web Service Discovery Based On The User’s Personalized Requirements

With the rapid development of web service, the same or similar functions of webservices have increased greatly, which give rise to a large number of irrelevant web serviceincluded in the list of returned services beyond user expectations when the user discoveriesweb services. Due to the different areas of expertise and different purpose of invokingservices, the user’s preference of service will also be different. The mechanism of webservice discovery is stricter in view of user personalization. So the research of web servicediscovery based on the user’s personalized requirements gets the attention of the industryand academia widely.But the research of current discovery of web service is in point of service provider tomake the implementation of the service, which loses sight of users’ personalization and therelationship between web services in the network. In view of the above shortcomings, thepaper mainly does two aspects of research:On the one hand, in order to solve the common problem that returned services can’tmeet the users’ personalization requirements in web service discovery, the paper presents ahybrid collaborative filtering algorithm for the recommendation of web services bycombining user-based and item-based collaborative filtering algorithm. The hybrid methodtakes into account the personalization of historical invocation information between usersand services and makes improvement on similarity calculation while adopting PearsonCorrelation Coefficient (PCC) to measure the similarity between two users or two services.Similar to product recommendation used in e-commerce sites and other fields, therecommendation of web service often take advantage of the history invocation informationof users and services (denotes as QoS, Quality of Service) to predict the unknown QoSvalue and then recommend web services to the active user with the optimal QoS.On the other hand, after the analysis of the process of enterprise business about webservice invoking, the paper implements a web service discovery and selection method ofvirtual social web services considering the relationship of competition or cooperationduring the web services. In the relationship of competition, firstly the paper calculates the similarity of text description information by analyzing the history of users invoking theservices and the user’s current requirements. Then based on the WSDL descriptionlanguage, the paper builds a tag tree of service operation and determines the similarity ofstructure between services after making improvement of tree edit distance algorithm. Wecould get a set list of services to meet user functional requirements to combine these twokinds of similarity. In the relationship of cooperation, after analyzing the characteristics ofservices, the method weighs the "communicative competence " of web service bycalculating the connection similarity between cooperative services in the network. Werecommend web services to the current user with good "communicative competence".Actually, the paper implements a virtual social mechanism of web service discovery andselection by analyzing the characteristics of services in these two kinds of relationship.