Recommendation System Based On Improved Collaborative Filtering Algorithm For Exercise Test

With the rapid development of information technology and network technology, the information on the Internet increases greatly, but it causes a big problem--information overload. In the face of these vast amounts of information, it is not easy for the users to search for the information or goods they are really interested in, so recommendation system appeared at the historic moment. It can provide personalized recommendation services which are based on different users’preference. It has been applied widely because of its unique advantage; especially in the field of Electronic Commerce it has achieved considerable success. The rapid development of Internet also makes the form of education changed undergone earth-shaking. Education is not only in the traditional classroom, but also can be carried out on the Internet. Online practice test is a kind of supplement of classroom learning, and it can be used as a platform of students’ autonomous learning. With more and more resources appeared, it is necessary to help the students choose the questions. This paper designs a collaborative filtering algorithm system, which is based on improved practice test recommendation. The system can recommend test questions in accordance with their learning conditions and personality characteristics for students.Firstly, the paper introduces the basic level in the domestic and foreign research recommendation system, narrates the main contents of the research, and the focuses and the significance of the research.Secondly, this article analyzes the collaborative filtering recommendation technology and classification in the personalized recommendation algorithm, describes the user-based collaborative filtering algorithm in detail, with including the comparison of similarity calculation method, and illustrates several standards to evaluate the quality of the recommendation:Thirdly, according to the actual demand, the paper designs the function of the system’s overall structure, as well as the recommendations algorithm adopted by the system. Innovation of this paper puts forward the calculation method of comprehensive similarity between users according to the personality characteristics of the users in the system. After researching the clustering algorithm, this article also puts forward the method based on Prim minimum spanning tree to improve the user clustering, and solves the problem of the initial clustering center randomly selected. Through students’clustering, it can effectively reduce the nearest neighbor query space in the algorithm, and shrink computing scale. The experimental results are verified by experimental data set, the algorithm can improve the efficiency and quality of recommendation system indeed, also solve the problems such as cold start and the sparse score matrix.This paper researches and designs a collaborative filtering recommendation algorithm based on Prim method to improve users’ clustering, and it is according to the algorithm implement a practice test recommendation system. The advantage of the system can recommend personalized questions service according to the characteristics of the student users. In the process of students’practice, the system not only promotes the students’ understanding and memory of subject knowledge, but also assists teachers to complete the objective assessment of student learning.