Cognitive Learning Analytics Using Assessment Data and Concept Map: A Framework-Based Approach for Sustainability of Programming Courses

Abstract

Students of initial level programming courses generally face difficulties while learning the programming concepts. The learning analytics studies, in these courses, are mostly anecdotal on the aspect of assessment as less or no attention is given to assess learning at various cognitive levels of specific concepts. Furthermore, the existing work reflects deficiencies in examining the effect of learners’ cognitive performance on subsequent stages of the course. This gap needs to be addressed by introducing more granular and methodical approaches of cognitive analysis for sustaining the programming courses effectively in computer science and associated disciplines. In this article, a framework-based approach is proposed for cognitive learning analytics on the concepts taught in initial level programming courses. The framework serves as a platform that provides structure to the concept data using the technique of concept mapping and examines learners’ cognitive propagation on related concepts using assessment data. Learners’ performance prediction has been examined on relatively higher-level programming concepts through the metrics established from the cognitive maps of learners, acquired by deploying the related layers of framework. Overall maximum prediction accuracy range obtained was 64.81% to 90.86%, which was better than the prediction accuracies presented in most of the related studies.