Report on the SIGKDD-2002 Panel The Perfect Data Mining Tool: Interactive or Automated Mihael Ankerst   View Paper (PDF)   Return to December 2002 - Volume 4, Issue 2 Abstract Commercial and academic data mining tools range from being fully automated to highly interactive. We will discuss the role of human involvement in the data mining process. On the one hand, providing interactivity/visualization enables domain knowledge transfer and the use of the human's perceptual capabilities. On the other hand, the vast amount of data to be mined today makes real-time interactivity hard to achieve and unnecessarily burdens the user to perform tasks that may be done automatically. Questions to be discussed include: What are the ideal roles of the computer and of the user in the data mining process? Which data mining methods (clustering, classification, association rules) can be improved by more human involvement? What kind of applications requires more human involvement? What kind of applications requires little or no human involvement? The panel was organized by Mihael Ankerst (The Boeing Company) and the participants were Surajit Chauduri (Microsoft Research), Georges Grinstein (University of Massachusetts Lowell and AnVil, Inc.), Jiawei Han (University of Illinois at Urbana-Champaign), and Gregory Piatetsky-Shapiro (KDnuggets). DiSC'03 © 2003 Association for Computing Machinery