In this paper, we outline a system for evaluating the performance of scientific research across a number of metrics. Our system is designed to cluster performance into a number of classes, classify performance on a metric with only data on other metrics, and to make possible the prediction of future performance on each metric. This study shows how data mining techniques can be used for business intelligence applications to provide a predictive analytic approach to the management of resources for scientific research.
Revised: June 27, 2012 |
Published: May 23, 2012
Citation
Bell E.B., E.J. Marshall, R.E. Hull, A.K. Fligg, A.P. Sanfilippo, D.S. Daly, and D.W. Engel. 2012.Classifying Scientific Performance on a Metric-by-Metric Basis. In Proceedings of the Twenty-Fifth International Florida Artificial Intelligence Research Society (FLAIRS) Conference, May 23-25, 2012, Marco Island, Florida, edited by GM Youndblood and PM McCarthy, 400-403. Palo Alto, California:AAAI Press. PNWD-SA-9767.