Font Size: a A A

Research On Ontology-Based Personalized Information Services Method

Posted on:2007-10-02Degree:MasterType:Thesis
Country:ChinaCandidate:Z W LiuFull Text:PDF
GTID:2178360212466753Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Study on personalized information services is a popular field right now, the existing study in this field often merely is limited to simple link selection and mechanical phrase matching, it can not understand the semantic information of user interest. Since its introduction in the early 1990s, Ontology theory attracts many foreign scholars'attention and related research in China has just begun in recent years. Ontology theory play an important role in the information organization aspect under the network environment.In order to achieve semantic relevant personalized information services, this paper has studied the user interest modelling, its expression and its renewal. This paper separately proposed the user interest discovery, the user interest expression as well as the user interest renewal method.Firstly, this paper introduces the Ontology theory as a semantic relevant personalized information services theoretical basis. It has constructed the specific domain ontology. Labelling the specific domain resources may produce the description database to express the specific domain resources.Secondly, in the user interest discovery aspect, it takes the user clicking flows as the metadata. It expresses the metadata as the vector in the vector space, proposes the integration method of K nearest neighbor algorithm and the support vector machine algorithm to discover user interest. In the user interest expression aspect, it expresses the user interest as a vector which is formed by specific domain Ontology category and its corresponding weight. This makes the semantic relevant recommendation much more easier.Thirdly, in the user interest renewal aspect, it takes the user's short-term interest and long-term interest respectively. Considering sliding window methods and gradually forgetting methods, a combined forgetting model is bring out to deal with the user interest updating issue.Finally, it put the Ontology based personalized information services study into the field of digital library applications. It introduces the Digital Library personalized information service system realization principle and the system overall flow, analyses the advantage of the new system, realizes a part of the...
Keywords/Search Tags:personalized information services, user interest model, Ontology, K-Nearest Neighbor (KNN), Support Vector Machine (SVM)
PDF Full Text Request
Related items