Information Use for Undergraduate Admission at Khon Kaen University and Students’ Learning Achievement Prediction Using Data Mining Technique

Abstract

The purpose of this study was to explore Khon Kaen University administrators’ use of information for admitting undergraduate students and to develop the students’ learning achievement prediction models using admission data mining. Data were collected for the study by conducting interviews from 12 admission administrators. The collected data were then selected to serve as predictive variables. Using decision tree approach, the researchers analyzed 39,696 items of integrated admission data of Khon Kaen University to predict the students’ learning achievement. The analysis of the data indicated that, in the admission consideration process, most of the administrators used test scores required by each subject field and personal information of the applicants. With regard to the development of the students’ learning achievement prediction model, it was found that the relationships of decision trees result in 5 prediction models with their accuracy of 63.32%. In addition, students whose learning achievement was found to be at the very high level were high school graduates from the Northeast and the South who were admitted to the University through the national central admission system.