Teaching Support System by Clustering Students According to Learning Styles

The student has different learning styles that indicate characteristics and skills of their learning. Assessment of student learning styles in the classroom will help the teacher know the cognitive skills of students. This research applies the hierarchical clustering technique for grouping students from similar learning styles in the classroom. In addition, the research investigates the parameters of hierarchical clustering that are appropriate for the learning style data and measures the efficiency of the proposed clustering method. Those parameters are the number of the cluster from 2 to 10 and two distance functions (Euclidean distance, Manhattan distance) and various linkage criteria (Average linkage, Complete linkage, Single linkage, Ward linkage, Median linkage). This research collects student learning styles into dataset using questionnaire for constructing student clusters. The experimental results show that grouping students into 7 clusters using the Euclidean distance function and the ward linkage criteria yields the highest efficiency in clustering (Error Sum of Squares, SSE = (569.457). Therefore, this method is suitable for the student learning style dataset collected in this research. The resulted clusters can help to identify the behaviors and learning skills of students that will encourage teachers to select appropriate methods and teaching tools.