Abstract:

Based on developments in management, team building has a significance influence on project management performance. This study describes the classroom team building using agglomerative clustering to improve individual and group performance in a computer science course.

Students features studied to determine:
(1) how well they know their classmates, and
(2) what process and strategies they use to communicate throughout the course.

Surveys and group interviews methods of inquiry were used to capture the efficiency of the group formation as well as the dynamic interaction within groups and the underlying factors that guided group communication and productivity. This study uses hierarchical clustering to form teams based on their acquaintance in an advanced computer science course at University of Houston during Summer 2019. The data analysis shows a clustering accuracy of 88.9%. Note that based on the pre-forming survey students who do not know anybody in class are clustered based on their programming skill and group preference. A post-forming group interview gives an insight into how different data mining algorithms might be used to enhance student performance through collaborative learning. These findings suggest that student formed teams can be an excellent solution to motivate students and to decrease the drop rates.

Keywords:

Educational Data Mining (EDM), Agglomerative Clustering, collaborative learning.