Assessing learners’ progress in a smart learning environment using bio-inspired clustering mechanism

Abstract

Learning Analytics systems can analyze and measure learners’ data to infer competence, meta-competence, and confidence measures. While catering to the needs of students, the Learning Analytics system also measures effectiveness and efficiency of the learning environment. These measures enable the Learning Analytics system to auto-configure and auto-customise itself to offer personalized instruction and optimal learning pathways to students. Such a Learning Analytics system can be classified a Smart Learning Environment, where learner engagement initiatives are auto-generated by the system itself. This paper proposes the Parallel Particle Swarm Optimization (PPSO) clustering as a mechanism to trigger learning engagement initiatives. Using PPSO, learners are clustered using similarity measures inferred from observed competence, meta-competence, and confidence values, in addition to effectiveness measures of instructional tools. A simulation study shows that the PPSO-based clustering is more optimal than Parallel K-means clustering.