Previous | Session 27 | Next
A. A. Suchkov, R. J. Hanisch, B. Margon (STScI)
We have applied ClassX, an oblique decision tree classifier optimized for astronomical analysis (McGLynn et al. ApJ, 616, December 1, 2004; Suchkov & Hanisch, ApJ, 612, 437, 2004) to the huge, homogeneous multicolor imaging data base of the Sloan Digital Sky Survey (SDSS), training the software on a subset of ~30,000 SDSS objects whose nature is precisely known via spectroscopy. We find that the software, using photometric data only, correctly classifies a very large fraction of the objects with existing SDSS spectra, both galactic and extragalactic. ClassX also accurately predicts the redshifts of galaxies and active galactic nuclei (AGN) in SDSS. Thus ClassX can probably address most of the highly diverse areas of astronomy addressed by SDSS, and should be especially powerful at accurately recognizing very rare objects solely through their photometric properties.
The surface density of AGNs selected by ClassX to r ~ 19 is in reasonable agreement with that quoted by SDSS. When ClassX is exercised on the photometric data fainter than the SDSS spectroscopic limit, the inferred surface density rises sharply, as expected. We discuss our measurement of the surface density as a function of apparent magnitude, which is unique due to the homogeneity of both SDSS and the classifier.
Finally, the ability of the classifier to accurately constrain the redshifts of huge numbers (ultimately ~ 107) of AGNs in the photometric data base provides new insights into the evolution of the AGN luminosity function with cosmic time. While our results are still very preliminary, it is clear that a very large population of relatively low luminosity AGNs appears in the 0.2 < z < 0.4 range.
Previous | Session 27 | Next
Bulletin of the American Astronomical Society, 36 5
© 2004. The American Astronomical Society.