The Research Of Resource Personalized Recommendation System Based On Education Website

With the rapid development of web technology, network education has been paid more and more attention. However, educational resources information overload brought difficulty to the user who selects effective resources. The personalized recommendation of education resources emerge as the times require. Personalized recommendation system can filter out the information which the users not interested in, and help users to extract the information they need from massive resources.This paper adopts the collaborative filtering recommendation method of personalized recommendation, and which carried out in-depth research. The K nearest neighbor algorithm and Rock clustering algorithm is improved, with Apriori association rules algorithm applied to the personalized recommendation system of educational resources. At the same time, according to the education website itself particularity, it introducs Ebbinghaus forgetting curve to enhance students’ learning effect. This recommendation based on the user access pattern similarity, find the neighbors of the target user, using the most similar neighbor user automatic recommended learning resources.Firstly, according to the classic K nearest neighbor algorithm, the principle was analyzed, found that more data, user similarity calculations required memory overhead is bigger, longer time. This paper proposes a set threshold of similarity and the weight of the method, the accuracy is more accurate, and the algorithm is more efficient.Secondly, according to the classic Rock clustering algorithm, this paper proposes similar weight and average neighborhood concept. It proposes one kind similarity calculation method based on the similar weight and average neighbor, narrowing the scope of user clustering, improving the accuracy of user clustering.Again, according to the students’ learning efficiency problem, this paper introducs the best memory time points of the Ebbinghaus forgetting curve, to provide students with the best study resources.The research of Education Resources Recommendation System based on the improved K nearest neighbor algorithm and Rock clustering algorithm, introduces the forgetting curve, improves the overall recommendation accuracy and coverage rate, which shorten the recommended time, and enhanced the students’ learning effect. The recommender system has been applied in the primary and secondary education website and obtained better practical effect.