Utility-Based Personalized Web Service Model

With the rapid development of electronic information technology, the information over Internet has undergone explosive growth. But the growth of information does not make the knowledge acquisition more convenient, it makes users to find the desirable knowledge in the vast information ocean even more difficultly. In order to solve such problem, Web personalized service technology emerges as the times require. Web personalized service technology is the combination of Web and data mining technology. Personalized service is a kind of "information to user’ service mode. With the application of Web mining technology, combined with Web content mining and Web domain ontology methods, the efficacy of Web personalized service technology can be improved.In recent years, under the joint efforts of the scholars, at home or abroad, Web personalized service technology has made significant progress. However, there are still several aspects should be improved, such as the "Cold-Start" problem, the better algorithm to update user interest model, personal recommendation algorithm and so on.To solve these problems, on the basis of the traditional personalized service model this paper build a utility-based personalized web service model. The utility theory is introduced in this paper, and based on it, we proposed a utility based user interests update algorithm. As to the "Cold-Start" problem, we use the abundant resources of third-part platform to build the initial user interest model to make the user get personalized service as soon as possible. At the meantime, to solve blemish of K-means, we proposed a collaborative based User-Interest double clustering algorithm.After the test of simulation experiment, this model can provide quality personalized service, satisfy the requirements of construction of personalized service of Web sites, and it has reality significance and valuable importance on academic research.