Retrieval Method Based On Fuzzy Object Of Annotations Clustering

With the development of Internet, the information of the network has an explosivegrowth. Since Tim Berners-Lee proposed the concept of Web2.0, users gradually transformedfrom receiving the information into a participant.Social annotation is one of the many applications of Web2.0. it’s a kind of meta-data thatthe mass users participated and shared. The mass users are not only the releaser of tags butalso the recipient of this information. With the continuous development of Web2.0, thesystem of the social annotation system adds many new services. The users described thecollected information in tags and shared them each other,which is the core services in manyservices.Social annotation is that the mass classify to the network resources based on their owninterested, and this annotation is without any criteria for the classification level. Although theemergence of the social annotation has brought many benefits for the users, it has many flawsfor the social annotation when users add labels arbitrarily.The searching method of this paper is based on the social annotation which is used toovercome the shortcoming of social annotation.The methed establishes the fuzzy object oftags. The thought come from probability model of information retrieval and the fuzzymathematics. The resources are very large on the network, so the search method of this paperuses fuzzy clustering methods to improve the efficiency of the queries.This paper establishesthe index tree for the fuzzy objects by fuzzy clustering. For the users’ query,The paper usesthe method of KNN query which returns to the users the first K results of the query. Thesearch method expansion tags for the queries of users according to users’ interested.Experimental results show that the researching methed of the paper proposed can bebetter meet user’s requirements.