Predicting Academic Achievement in Fundamentals of Thermodynamics using Supervised Machine Learning Techniques

Supervised machine learning techniques were used to answer the central question: does excelling in reading quizzes, a good predictor of accurately predicting the passing rate in MEEN241 Fundamentals of Thermodynamics? Class assignments such as reading quizzes (RQ), quizzes (Q) and home-works (HW), Tests (T), and midterm (MT) were designed. The predictor variables analyzed are High GPA (>3.0), RQ, Q, and passing PHYS241 General Physics and these were used to develop two classifiers: CART with cross-validation and Random Forest. The CART and the Random Forest models identified Q, and Q and RQ, respectively as the best predictor, although quiz and reading quiz accounted respectively, for only 5% and 15% of the total weight. This suggests that students who devote the time and effort into doing the reading assignments and subsequently passing both the RQ and Q are likely to expend similar effort and time in other class assignments and preparation towards tests and examinations.