A thesis submitted in partial fulfillment of the requirement for the degree of Master of Science, School of Engineering and Technology

Thesis (M.Sc.) - Asian Institute of Technology, 1994

Internet is one of the fastest growing human-constructed phenomena in history. The explosive growth of the Internet has brought with it corresponding growth in the amount of information available. A fundamental problem confronting users of such network is how to cluster people by shared interest to support interpersonal resources discovery, based on electronic mail communication patterns. This research is concerned with the clustering of nodes based on electronic mail communication. Such clusters will be useful to discover shared interests between the users, and between users and resources. We have developed an approach based on self organizing maps to partition an arbitrary network of communication into clusters representing shared interests. The method has been compared with graph partition methods.