Prediction of the number of students who will take themake-up exam using decision trees and artificial neural networks

Make-up exam is a type of exam held to give a second chance to the student who has failed a course as a result of the final exam to pass the course. However, for various reasons, a significant number of students do not exercise their right and do not take this exam. In this thesis, it is aimed to prevent unnecessary waste of resources, energy and labor, such as printing excessive exam papers, opening redundant exam halls and assigning invigilators unnecessarily, by predicting the number of students who will not take the exam. In this thesis, an attempt was made to predict the number of students who will not take the exam in order to prevent unnecessary use of resources and workforce. For the study, the characteristics of each student's non-personal data (gender, overall weighted grade point average, semester grade average, course midterm exam grade, final exam grade, etc.) were determined. Considering these determined features, data sets for some courses were created and applications were developed with Artificial Neural Networks and Decision Tree algorithms using these data sets. When the results obtained from these two classification methods are compared; The accuracy rate was obtained as 88.70 with the Artificial Neural Network and 87.96 with the Decision Tree algorithm.