Clustering multiple mix data type for automatic grouping of student system

A kind of learning process that following some instructions involving students work in groups to achieve goals called cooperative learning. Cooperative learning can help students to be more confident to tell opinions and increase their knowledge more effective. In cooperative learning, group formation of students is important. Preferably, teacher determinate group learning based on result of teacher observation. But, it is not easy and takes long time to get the best heterogeneous group formation. More students, harder. The easiest way to determine group of students is randomly, but we can't make sure the group is heterogeneous. So, cooperative learning process become inefficient. Hence, we need a system that capable to determine group formation automatically. In this paper, we present the experiment of cooperative learning. Group formation determined automatically by system based on dissimilarity between student of mix data types, by teacher and by student manually. After we did experiment in 3 classes, we analyzed experiment data use F-test. The result is system can determined heterogeneous group as good as teacher grouping.