A self-assessment system using machine learning for empowering graduate students

Abstract

This study presents the development of a self-assessment system that employs machine learning techniques to predict graduate students' likelihood of completing their studies within the designated program duration. Data from 33 graduate students were collected through a structured questionnaire covering 38 influencing factors. The dataset was preprocessed and expanded using the SMOTE technique to enhance prediction accuracy. Two primary models were implemented: Logistic Regression was used to classify whether a student would graduate on time, achieving an accuracy of 90%, while the Random Forest technique was used to predict the expected duration of study with 84% accuracy, a Mean Absolute Error (MAE) of 4.52%, and a Root Mean Squared Error (RMSE) of 4.93%. The system was developed using Python and Visual Studio Code and features a user interface for entering personal attributes and displaying prediction results. The system serves as a practical tool for students in planning their academic paths and for institutions seeking data-driven strategies to improve graduate outcomes. It also contributes to the growing body of research in educational data mining and self-assessment technologies.