Simulated stellar classification combining the minimum distance method with a maximum likelihood procedure
| Autor: | V. Malyuto |
|---|---|
| Rok vydání: | 2002 |
| Předmět: |
Physics
business.industry Template matching Photometric system Astronomy and Astrophysics Pattern recognition Linear interpolation Stellar classification Spectral line Stars Template Space and Planetary Science Astrophysics::Solar and Stellar Astrophysics Artificial intelligence business Instrumentation Random variable |
| Zdroj: | New Astronomy. 7:461-470 |
| ISSN: | 1384-1076 |
| DOI: | 10.1016/s1384-1076(02)00170-7 |
| Popis: | A modification of the minimum distance method (or template matching) of classification has been tested where a maximum likelihood procedure is added. This two-stage combined method provides continuous classification with smooth linear interpolation. The present analysis is based on the synthetic color indices of the Vilnius photometric system calculated from the Kurucz synthetic spectra for F–G–K stars. We have simulated ‘observed’ color indices for several models by combining synthetic color indices with random variables. Our classification method has been applied to the ‘observed’ stars with the use of the synthetic color indices as templates. The classification accuracies for different photometric accuracies have been estimated with the use of less dense and more dense grids of templates; the models included and not included into the grids have been considered. The conclusion is that our method is able to provide continuous photometric classification of good accuracy but the grids of templates should be dense enough, especially at higher photometric accuracies. The method may be applied to other photometric and spectral classification systems. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect