Research On Personalized Recommendation System Based On Ontology Of User Interest Model

With the development and the popularization of information technology in twenty-first century, web site has gradually become an indispensable part of people’s life, it has brought great convenience to people’s work and study. Along with the increasing of the website, the sharp increase of the amount of information, the popular "one-size-fits-all" operation mode of the traditional web site has been unable to meet the user’s personalized needs of different background. How to show itself in all kinds of sites, to prevent the old user passes, to attract new users access at the same time, it becomes a problem for the website system operators. Personalized information recommendation system is emerged as the times require in this background, and quickly become a research hotspot in recent years.From the initial technology, such as collaborative filtering, based on web mining, domain ontology, to the latest technology, such as artificial intelligence, distributed processing, the personalized recommendation has made rapid development. At present, most of the personalized collaborative filtering recommendation based on ontology research also has shortcomings, for example, ignoring the importance of the relationship between ontology concepts, ontology user interest model updating is not complete, user interest ontology concept can not be forgotten with the passage of time.In order to solve these defects, this thesis puts forward a user interest model based on ontology, the thesis combines activated extension theory to adjust the interest model updating algorithm and introduces the user interest degree forgotten formula. When the user’s interest changes, the personalized recommendation system can make thorough update for the whole network of user interest model concept node according to the activated extension theory. In addition, by calculating the similarity between user interest model, instead of the user for different things score, produced a synergistic personalized recommendation list. Finally, the thesis cites the Movie Ontology as the film of domain ontology that developed, adjusted and maintained in informatics laboratory at the Zurich university, using the MovieLens and IMDB movie data set as experimental data set, design and implement the movie personalized recommendation system based on ontology of user interest model. When the user login successful, the system will collaborative filtering recommendation based on user ontology and preference data, generate personalized recommendation list to the user.In the whole, the current application of recommendation system based on personalized ontology film is still in the research stage, the open source of research project for web users recommendation system is relatively lack of both at home and abroad, this thesis designed and implemented movie personalized recommendation system based on ontology of user interest model, it is a good try on this field.