Use of a weighted application form for demographic data an the attribute selection technique to evaluate freshmen's ability to succeed (Thai)

Abstract

This study involved the use of a weighted application blank (WAB) technique to compare 2 attribute selection techniques; 1) correlation-based feature selection and 2) consistency-based subset evaluation. The purpose of this study was to evaluate freshmenûs ability to succeed from demographic data. The results are a proposed WAB rule by median with consistency-based subset evaluation by best-first search method, giving a high performance accuracy rate of 72% and reducing the number of attributes by up to 66.66%. This WAB can be applied to create a model with guidelines to equalize new students and use the results to improve courses or curricula in the future.