Discovering group interaction patterns in a teachers professional community

Abstract

Managers and group leaders of a community may not have enough information to help them direct the community development. In this paper, we adopt social network analysis, graph theory and data mining techniques to analyze groups in a community at their different project periods. Each group is presented by several attributes that are computed from social network analysis and graph theory to represent various interactions of the group members. The evolution of these attributes enables the discovery of the interaction patterns of different groups in their life cycles. With these discoveries, team leaders can obtain concise information about their teams' performance, and community managers can cluster groups to capture stereotypes of virtual teams in the community. Such knowledge will benefit group development in a community.