Web Users Based On Fuzzy Multisets And Page Clustering Model And Algorithm Research

It becomes one of the key factors to the web system that how to provide personalized and flexible service to different users in web enviroment.It also affects whether web system attractive and application.Clustering of web users and web pages is the basis of personalized service.The practicality of research and establishment on new model and algorithms of web users and web pages clustering are significative.The main results of the thesis are as follows:Firstly,a model of web users and web pages clustering based on fuzzy-multisets is proposed.The theory of fuzzy-multisets is combined to the clustering of web users and web pages.It makes use of multiple attributes to describe the objects of web users and web pages. After characterizing the objects of clustering by fuzzy-multisets,the model of clustering is constructed.The algorithms of HCFM,FCFM and CAFM are described in the thesis.Secondly,a method of web users and web pages similar measurement is proposed in this thesis.The similarity of web users and web pages is described by the distance between two fuzzy-multisets.A measurement of web users and web pages is put forward.Thirdly,a pretreatment model which transforms the weblog data to clustering object set is constructde.Data collection,data cleaning,data transformation and data standardized are processed.A good foundation has been set up for the web user and page clustering.In the end,a prototype system of web users and web pages is implemented.In this thesis,a clustering prototype system is designed by means of the JBuilder and ACCESS technologies. Testing and analysing the model and algorithm of clustering on basis of the educational web log data.The feasibility and correctness of the clustering model and algorithm are validated by experimentation.