The Design And Implementation Of Personalized Recommendation System For Digital Media Based On User Context

Today, personalized recommendation system has been applied in many fields, such as online products recommendation, music recommendation, etc. Collaborative filtering is one of the most widely used technology in personalized recommendation domain, but most of these systems do not take into account the user context to the impact of recommended results, or the recommendations do not change with the update of the users interest preferences.In this paper, on the basis of the personalized recommendation system which was implemented using the traditional collaborative filtering technology, we introduce the user's context information into the recommendation system. Under the in-depth research to context-aware computing and ontology-related theoretical knowledge, the paper analyzes and implements the system for digital media recommendation based on the user-context.The work accomplished in this paper includes:Firstly, based on the preliminary requirements analysis it determine and point out the concrete contents of the user-context information in the system, and give the solution to retrieve the user's context information from the web pages. Secondly, we construct the domain ontology for digital media in the system and model the user's context information using the ontology modelling techniques, implement the automatic creation and expansion of the ontology individual instances for the various concepts in the system. Thirdly, we define the representation of the user interest preferences and build the user interest preferences for every user in the system based on the ontology query and reasoning capability. Finally, based on the user preferences established and the user's current context information, the system computes a number of recommended results and presents the results on the front page.Overall, it is still in the research stage about how to use the context-aware computing and ontology theory in the actual recommender system. And most of the research project is aimed at mobile service recommendation; the recommender system for web users is relatively less. The system implemented in this paper is an attempt on the context-aware and ontology research domain.