Enhancing student clustering to generate adaptive metacognitive instructions in learning system for vocational high school
Indriana Hidayah, Teguh Bharata Adji, Noor Akhmad Setiawan, Norliza Abd Rahman · International Journal of Innovation and Learning · 2018
Adaptivity in learning systems depends on accuracy of learner modelling. Specifically, for generating cluster-based instructions, quality of student clustering is critical. Studies on student clustering are abundant; however, a system for clustering metacognitive that considers proper analysis technique of Likert-scaled dataset is unavailable. This article proposes a student clustering method which uses a new Likert scale analysis. It is performed on a dataset collected from 81 students of a vocational high school. The performance was compared to previous methods; enhancement is shown by higher silhouette-index and strong correlation with work readiness score. To evaluate the clustering effectiveness, it is implemented on an e-learning system to generate adaptive instructions. The e-learning system is a supplement for fundamental programming course in the school. The t-test result shows that learning gain of experiment group is significantly higher that of the control group. Therefore, the proposed method is effective in improving students' learning quality.