Applying Stacked Generalization with the Difference of Truth and Falsity Data to Predict Student’s Performance

Abstract

In this paper, the difference of truth and falsity data is applied to stacked generalization. The truth and falsity data are created in level 0 of stacked generalization. In level 1, the truth data produced from level 0 together with the difference of truth and falsity data are trained. This proposed approach is used to predict performance of undergraduate students in the Department of Computer Science, Ramkhamhaeng University, Thailand. It is found that the proposed technique provides better accuracy result than the existing stacked generalization techniques for the prediction of student’s performance.