Support System to Pbkp_redict Student Dropout in Universities

Abstract:

The objective of this research was to provide a pbkp_rediction system for the possibility of student dropout at the Pontificia Universidad Católica del Ecuador Sede Ibarra. It is applied research with a mixed approach. It was developed in two phases. In the first phase, the KDD methodology and the Scikit-Learn tool were applied to select the best pbkp_rediction algorithm (KNN, Decision Tree, Random Forest, SVM, and Neural Network). In the second phase, the information system was built to make use of the model obtained in the first phase, where users will be able to consult the possibility of risks of academic dropout of students. Technologies such as Django, Python, HTML, JavaScript, and MySQL, among others, were used in this study. The results show an information system that allows consultation by the student, by the level of schooling, or by subject, based on neural networks that provide an accuracy of 92%.

Año de publicación:

2023

Keywords:

• learning analytics
• Student dropout
• data analytics
• Data Mining
• pbkp_redictive models

Fuente:

scopus