Student Performulator: A Tracking System for Performance Measurement of Undergraduates Using Convolutional Neural Networking Model

Abstract

At present, all the higher educational institutions attempt to ensure their quality in various aspects to meet the local and international standards. In this context, however, most university students lack knowledge of their progress in the degree programme they are enrolled in. Therefore, there is a need for a tracking system in universities to understand and inform the relevant stakeholders of the undergraduate performance, enabling them to make correct decisions. Hence, the main objectives of this study are to provide the management a way to monitor and manage the students’ progress and to update the students of their status according to pre-determined zones. A dataset containing 27 attributes and 9504 records was collected, considering the undergraduates from different levels in selected faculties in the University of Jaffna as the sample of this study. The researchers used the statistical technique of Chi-Square test to select the most appropriate predictor variables for student status prediction. Then, the Convolutional Neural Network (CNN) analysis was performed on Python programming with Pandas and NumPy libraries and Excel functions were carried out for the purpose of transforming categorical variables into numerical values. The experimental results show that the proposed model yielded best results up to 81.4 %. Researchers believe that this model could be applied by different faculties in other universities and future studies could be extended by improving this model.