Multi-Data Mining for Understanding Leadership Behavior

Authors: Naohiro Matsumura, Yoshihiro Sasaki, Osaka University

We propose an approach for understanding leadership behaviors in dot-jp, a non-profit organizations, by analyzing heterogeneous multi-data composed of questionnaire and mailing list archives. Reliabilities of leaders are obtained from the questionnaire, and the human networks are extracted from the mailing list archives. By integrating those two results, we discovered that leaders should receive messages from other people as well as send messages to constract reliable relationships.