Predicting Academic Performance From Admission Scores and Application Data – A Case Study
During the admissions process at the University of Tartu’s Master's programme, students get an admission score, which is calculated based on their motivation letter and their Bachelor’s Grade Point Average (GPA). Ideally, the admission score should be predictive of student’s academic performance during their Master's studies. Hence, one would expect that the higher the admission score is, the better is the academic performance of the candidate once they join the Master's program. However, this hypothesis has not been tested in the context of University of Tartu's admission process. In this thesis, we evaluate the efficiency of the admission score and produce more accurate and reliable student’s academic performance prediction model by analysing the application documents using data mining techniques.
Graduation Thesis language
Graduation Thesis type
Master - Software Engineering
Marlon Dumas, Irene-Angelica Chounta