Research on Three-Layer Collaborative Filtering Recommendation for Online Judge

Online Judge is widespread for the undergraduate to study programming. The users usually feel confused while locating the problems they prefer from the massive ones. This paper proposes a specialized recommendation model for the online judge systems in order to present the alternative problems to the users which they may be interested in potentially. In this model, a three-level collaborative filtering recommendation method is referred to and redesigned catering for the specific interaction mode of Online Judge. This method is described in detail in this paper and implemented in our demo system which demonstrates its availability.