Forgetting punished recommendations for MOOC

Abstract

Prerequisite inadequacy tends to cause more drop-out of MOOC. Recommendation is an effective method of learning intervene. Existing recommendation for MOOC is mainly for subsequent learning objects that have not been learned before. This paper proposes a solution called Forgetting-punished MOOC Recommendation (FMR). FMR combines the forgetting effect on learning score as a main feature for recommendation. It provides Prerequisite Recommendation (PR) for the unqualified learning objects and Subsequent Recommendation (SR) for the qualified objects. Experiments verify the accuracy improvement of PR and SR.