Quantum Course Prophet: Quantum Machine Learning for Predicting Course Failures: A Case Study on Numerical Methods

This study delves into the application of Quantum Machine Learning to predict student course failures based on their performance in prerequisite courses. Specifically, we adopt Quantum-enhanced Support Vector Machines to develop an intelligent system called "Quantum Course Prophet" to forecast whether a student might fail the numerical methods course. We used a dataset comprising 103 examples from the academic histories of students enrolled in the Systems Engineering bachelor's degree program at the University of Cordoba in Colombia. Notably, the Numerical Methods course involves 10 prerequisite courses in the latest version of the curriculum. For each course, we included in every student's example the highest final grade, the lowest one, and the number of times the student has enrolled in the prerequisite course. Consequently, each student is represented by 33 independent variables, with the target variable indicating whether the student is likely to fail the Numerical Methods course. To deal with memory constraints, Non-Negative Matrix Factorization is employed to reduce the dimensions of each example. Following dimension reduction, each variable is rescaled to standardize the maximum value to 1. The dimension of the input space is determined through 10-fold cross-validation, resulting in a seven-dimensional input space. This approach yields a mean accuracy of 73.82%, precision of 70%, recall of 61.5%, and a harmonic mean of 64.44%. This study underscores the potential of quantum machine learning as a viable alternative for addressing real-world problems in the future.