Research On User Model Based On Tag In Personalized Service

With the development of the technology and popularity of the network, people enter the information explosion era, personalized information services become the research focus, and as the core of personalized service, user model's quality directly affects the quality of personalized service.There are so many methods about user model, and the model improved the quality of personalized service to a certain extent. But these methods are basically extracting the user interest features from the users who have browsed the documents. The method not only to waste a lot of time to extract keywords from the documents, and, different users who browsing the same documents, the demand for content point of view may be different, but the system will extract the same keywords for different user, this will can't highlight the meaning of"personality", and will reduce the quality of personalized service. With the development of Web technology, label (Tag) service is proposed, because of tag's randomness and simplicity, more and more users like use tags, and tag has been widely used in the Internet.In this paper, for the shortcomings of the current user model and the extensive use of tags, we research user model based on tag.Firstly, through some experiments prove that the tags which user using contains stable interest and the characteristics about tags distribution;Then, presents a weighted tree from coarse to fine granularity describes the user model, To improve the timeliness of service establish weighted frequent item-set table for often used together tags for each user, and reasonable treatment of customers"own"words.Again, calculation tags'similarity based on WordNet, and use the clustering algorithm to create the user model, the paper introduced how to update the user model and the process of realization of personalized service with this model.Finally, select the data set filter simulation and analysis the experimental results.The model avoids the complex process of extracting key words and we use tags from users'bookmarks to expresses their interests from the user's point of view.