Multivariate Predictive Modelling of Mathematics Semestral Grade via Bayesian Networks Machine Learning Algorithm
Purpose – This study pertains to the novel use of Bayesian Networks to elucidate the interplay between multivariate factors of demographics, personality types, mathematics anxiety, and study habits in predicting the mathematics grades of college students. The research tried to uncover the combination of predictor variables that would likely explain the failure of students across different math courses.
Method – The methodology follows a survey of 1200 DLSU-D students across the seven colleges of the university. Personality types were evaluated using the Myers-Briggs Type Indicator (MBTI), and mathematics anxiety was measured using the Abbreviated Math Anxiety Scale (AMAS). The machine learning implemented the framework of Probabilistic Graphical Models in Python (PGMPy) for data visualization and analyses. Predictions of possible grades were summarized, and the full Bayesian Network was established.
Results – Bayesian analyses have shown that the chances of failing a math subject are generally low for each year level. Personality variables conclude that college students with analyst roles have a higher probability of having a perfect 4.00 grade in a math subject than in an explorer role. Predicting the chances of failure between having or not having math anxiety seems almost no significant difference.
Conclusion – The subject, anxiety, consistency, and enjoyment variables are minimally enough to infer the probability of mathematics grades; hence all other variables can be ignored. There was a total of 8 math subjects with predicted probabilities of failing students over a total of 13.
Recommendations – Researchers recommend the use of other probabilistic graphical models aside from Bayesian Networks to verify and compare the joint probabilities between the variables of the study.
Research Implications – Providing comprehensive insights to properly accommodate at-risk students in each math subject will greatly help mathematics professors recalibrate their attention and teaching strategies.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.