INITIALIZING STUDENT MODELS USING DEMPSTER-SHAFER THEORY

摘要

Students bring preconceptions into learning situations. If their prior understanding is not engaged, they may fail to grasp new concepts. We present a pretest scoring algorithm to initialize a student model. Initializing student a model can reveal student background knowledge with respect to the domain content. Dempster-Shafer theory was applied to cope with uncertainty factors in using a test which included incomplete data and students' unsolicited behaviors. Many concepts may be needed to answer one question correctly; therefore questions of the pretest can be used as sensors that collect evidence of student preconceptions. The system is stand-alone and easily employable by any intelligent tutoring system. The applicability of our approach was demonstrated by using the scoring result with an overlay student model.