Using the basis of a knowledge space for determining the fringe of a knowledge state

Doignon and Falmagne have developed the concept of the fringe of a knowledge stale based upon the neighborhood of knowledge states. This concept is useful for probabilistic assessment of knowledge and for the suggestion of training problems in an intelligent tutoring system. In this paper, a generalized definition of the fringe of knowledge states is introduced. An efficient procedure is presented which computes the fringe of knowledge slates using the basis of a knowledge space. In a simulation study, this procedure is compared with the approach described by Doignon and Falmagne. (C) 1997 Academic Press.