Course Prophet: A Machine Learning-Based System for Early Prediction of Student Failure in Numerical Methods Course in the Bachelor’s Degree in Engineering at the University of Córdoba, Colombia

Abstract

In this paper, we address the following research question: Is feasibility to use an artificial intelligence system to predict the risk of student failure in a course based solely on their performance in prerequisite courses? Adopting a machine learning-based quantitative approach, we implement Course Prophet, the prototype of a predictive system that maps the input variables representing student performance to the target variable, i.e., the risk of course failure. We evaluate multiple machine learning methods and find that the Gaussian process with Matern kernel outperforms other methods, achieving the highest accuracy and a favorable trade-off between precision and recall. We conduct this research in the context of the students pursuing a Bachelor’s degree in Systems Engineering at the University of Córdoba, Colombia. In this context, we focus on predicting the risk of failing the Numerical Methods course. In conclusion, the main contribution of this research is the development of Course Prophet, providing an efficient and accurate tool for predicting student failure in the Numerical Methods course based on their academic history in prerequisite courses.

Course Prophet: A System for Predicting Course Failures with Machine Learning: A Numerical Methods Case Study

“…Notably, this latter work was dedicated to the Numerical Methods course within the bachelor's degree program of Systems Engineering at the University of Córdoba in Colombia. In the present paper, we have achieved enhanced results in comparison to the findings published in [6].…”

“…• artificial neural networks or multilayer perceptron [1][2][3][4][5][6] ensemble methods with different classification methods [1,4];…”

“…• random forest [2][3][4]6], • gradient boosting [3]; extreme gradient boosting (XGBoost) [3,4,6]; • variants of gradient boosting [3,7], namely CatBoost [9] and LightGBM [10]; and • Gaussian processes for classification [6].…”

“…In contrast to the studies mentioned earlier, in [3,6], the goal was to forecast students' risk of failure prior to the course beginning, using their grades in prerequisite courses. In [3], grades were represented as ordinal data rather than quantitative data, and students' ages were also used for prediction in addition to grades.…”

“…In [3], grades were represented as ordinal data rather than quantitative data, and students' ages were also used for prediction in addition to grades. In [6], the predictive input variables encompassed scores in the admission test and grades in prerequisite courses. Notably, this latter work was dedicated to the Numerical Methods course within the bachelor's degree program of Systems Engineering at the University of Córdoba in Colombia.…”