The Applied Research of Collaborative Filtering Algorithm on ACM Online Judge Recommend System
Sun Qua · Computer and Information Technology · 2015
ACM online judge System based on the number of problems as much as problems of blind students, this paper analyzes the project- based and user- based collaborative filtering algorithms are used to calculate the cosine similarity between users(or projects) similarity in the calculation of individual users when recommended, in order to discover the current user preferences, in order to discover the current user preferences, so the sooner the problem- solving by weighting recorded smaller weights. Impact and compare the training set through experiments accounting data set ratio, the number of problem users AC, recommended to the user the number of factors such as the subject of recommendation quality, the results showed that: OJ on our system,the user- based than the project- based recommendation algorithm better quality of recommendation. Finally, the recommendation algorithm deployed in the OJ systems, and user generated title recommend personalized recommendations designed to help students better training, and gradually improve the ability of university students using computer program to analyze and solve problems.