Forecasting of University Students' Turkish Language Course Exam Results According to Their Exam Preparation Levels

Abstract

There are many algorithms and software in the field of forecasting. These methods are also used in education. There are studies on forecasting of student achievement in education. Forecasting of academic achievement of university students is important in terms of seeing possible situations. In this study, the achievement of the students towards Turkish Language course was forecasting with data mining methods. 160 student data, in a state university, were included in the study. For the data obtained, prediction models developed by DecisionStump, RandomTree, RandomForest, REPTree and M5P methods were created and compared with each other. 10-fold cross-validation method was used in the separation of data for training and test. In models, it will affect the student's passing grade; program, type of OSS entrance, OSS entrance score, OSS entrance rankings, the previous semester grade point average, midterm exam grade, studying status, the current study and how many points are expected from the exam, how the exam passed and final exam score was taken into consideration. Among the models, it was seen that the model established with best results with 10.16 mean absolute error and 0.72 correlation coefficient. As a result of the study, it is thought that students can take precautions by predicting the passing grade.

Keywords

• data mining
• forecasting
• turkish language
• university students