The Research On Information Service Of Personalization Based On Ontology

With the development of Internet, volumes of text documents have emerged in recent years, which lead it difficulty to obtain useful information by users. How to actively provide information service for users that can help them effectively retrieve the information they need, namely Information Service of Personalization, has become an important trend of information service. Mining, analyzing and studying web log is an important way by which Personalization can be realized. We can either understand reflections and motivations of users or identify latent users of e-commerce, comprehend interests of final users, and improve qualities of information service via mining amount of log information and analyzing rules of log. So web log mining is one of the most important topics in Information Service of Personalization. Since this research in China is at its early stage, and there exist many problems that we will study in this paper. First, we introduce the background and current development of automatic Personalization and discuss the proposal, content and target of our research. Second, we design an automatic Personalization Model based on Web Usage Mining and research main aspects of it such as preprocess of web log, process of Data Mining and algorithm of recommendation, then lay emphasis on process of Data Mining, present Association Rule Mining algorithm and increment updating algorithm. The results of the experiments are discussed. Finally, we present a general framework for fully integrating domain Ontology with Web Usage Mining and Personalization processes, give algorithm of recommendation and relativity algorithm of recommendation, and compare with formerly algorithm of recommendation by lab. The main results of this paper are as follows: present an automatic Personalization Model based on Web Usage Mining; give an Association Rule increment updating algorithm fits for Web Usage Mining; design an algorithm of recommendation of automatic Personalization based on Ontology and an algorithm of recommendation based on the relativity of domains.