Research On Folksonomy Based Web Services Discovery

To improve the efficiency of service discovery, identification and third party classification system was introduced into UDDI 2.0. But it is still incapable of exact search, because of its massive disadvantages. In traditional Web Service discovery model, Web Service description is totally dependent on Wed Service provider, which makes the incomplete and unreadable. Because of the broad participation of users, Folksonomy has great flexibility and many advantages on web resources management.A service discovery model based on Folksonomy is proposed to solve the high cost and difficult maintenance problem of traditional classification system. After using a service, a user can tag it with semantic assertion. When the number of tags reaches a certain scale, the whole Folksonomy system will form a comparatively complete classification system. It will be perfected with the participation of users, and finally realize exact search. By adding tags, users can participate in the process of Web Service description, making it more comprehensive and complete so that the problems of the old pattern can be avoided. In solution to the shortages of Folksonomy itself, some improvement suggestions systems are made.After the analysis of tagging system's network characteristics, a hierarchical tag clustering algorithm is proposed. By hierarchical tag clustering, the correlations between tags are analyzed. Then Ontology structure is derived from tag clusters for semantic parsing and Web Service classification storage.This paper focuses on the research of Folksonomy based service discovery. It carries out initial explorations of the combination of folksonomy and service discovery mechanism, including research on the model of service discovery and the mechanisms of tag clustering, service clustering and Ontology deriving.