[Previous] | [Session 10] | [Next]
R. M. Humphreys, G. Karypis, M. Hasan (University of Minnesota), J. Kriessler (Efficient Channel Coding, Inc.), S. C. Odewahn (Arizona State Univ.)
Image classification and pattern recognition is an important and challenging problem for the efficient analysis of large astronomical databases and will become even more important with the development of the National Virtual Observatory. Neural networks and decision trees have already been successfully applied to star-galaxy or stellar-nonstellar image separation. But for many astrophysical problems such as large-scale structure, galaxy formation, and evolution, we need the morphological type of the galaxy. We report the VERY promising results of our first experiments with different classification algorithms including decision trees, K-nearest neighbor and the new support vector machines.
Our results will be applicable to other digitized and digital sky surveys that will be part of the NVO.
The APS Project and this investigation are supported by NASA's Applied Information Systems Research Program.