Predicting Academic Achievement with Machine Learning Methods

Abstract

In spite of the projects and physical investments implemented in the field of education in our country in recent years, the successes of the students in the national and international examinations has not increased at the required level.Particularly, this failure situation is much more serious for Turkish and Mathematics lessons. There are many family, demographic, school based factors that affect the academic achievement of students among middle school students. However, until now, it has not been clearly stated which factors have priority, directly influence success. In this study, a questionnaire of 24 questions prepared according to the determined criteria was applied to 6th, 7th and 8th grade students of secondary school. The scores and grades of Turkish, Mathematics lessons and the overall success averages were estimated by using regression / classification machine learning models on the data obtained from the survey results. According to the experimental results obtained, the prediction of the student's grades has been successfully implemented with machine learning methods.

Keywords

• Course success average estimation
• Feature selection
• Regression
• Classification