Gaze-Based Prediction of Students’ Math Difficulties - A Time Dynamic Machine Learning Approach to Enable Early Individual Assistance

Abstract Graphical differentiation substantially promotes the understanding of basic concepts of calculus. At the same time, such tasks are challenging for many students. To be able to support students in complex tasks during the solution process, it is important to recognise if students are having difficulties and, if so, to identify precisely what the difficulty is. However, previous research was only able to identify correct and incorrect solution processes but not the specific difficulties of students. In our study, the eye tracking data of 143 students reveal how students’ math tasks in context of graphically differentiation can be predicted. Retrospective-think-aloud (RTA) protocols were used to identify students’ difficulties and three machine learning algorithms, k-nearest neighbours (KNN), Random Forest and support vector machine (SVM), provide an accurate prediction for this multiclass problem. The prediction results improve with increasing processing time and achieve already good values long before the solution process of the tasks is completed. Our results show that certain student difficulties can be detected very early during the task solution process. Although this approach has been demonstrated for a sub-area of calculus, it is transferable to other fields of the STEM domain and therefore has much wider scope.