Research Of The Personalized Recommendation Of The Learning Content Based On Knowledge Point

With the rapid development of Internet, the global demand for lifelong education and extensively carrying out the modern distance education, E-Learning is becoming one of the important application of Internet whose main feature is the asynchronous education Meanwhile, providing personalized service for user has become more and more urgent to E-Learning system. According to customer's characteristics such as interest, hobby, the level of understanding and so on, the personalized service would provide different services for different users. Distributed Internet resources and the environment can provide strong technical support for the distance education and knowledge services in the way of accessing to the knowledge and collaborative leaning approach. Internet learning resources is becoming thrived, but there are still exist some problems in the application of E-Learning. For example, the personalized learning under dynamic Web heterogeneous environment, granularity of recommendation resource and so on. In light of these problems, this paper studies deeply in the aspects of constructing user model model, the user intent matrix model, and algorithms of personalized recommendations and so on, and proposes an improved personalized recommendation algorithm.The main contents of the research include: constructing the personalized user model; research and measure of the user intent matrix; studying and improving the personalized recommendation algorithm; designing the recommendation system for Computer Network.1) Studying the representation methods of vector model based on key words, user-project evaluation matrix and ontology. This paper chooses the method including ontology to represent the user model. This method can represent complex basic information and extensional information of the users, and it's easy to update users'model and find the users'potential interests by the reasoning mechanism of the ontology.2) Studying the effect of users'behavior in their profile; Finding the measure of users'interests by studying in the aspects of resource type, feedback results of testing questionnaire, resources saved by users, kept resource, access frequency of resource and users'detention time in resource.3) Improving the current recommendation algorithm at content learning by studying and analyzing current recommendation algorithms; solving the sparse matrix problem effectually by substituting user-type evaluation matrix for user-knowledge evaluation matrix, and reducing the sparseness of data; demonstrating the effect of the improved algorithm by experiment.4) Constructing the ontology library of Computer Network; designing a content learning personalized recommendation system based on knowledge using the improved recommendation algorithm.