2004/03/10 by Ranjan Gupta, Harinder P. Singh, K. Volk +1 · 1 citation
Engineering · Physics and Astronomy · #Artificial neural network #Astronomical Observations and Instrumentation #Astronomy and Astrophysical Research #Astrophysics and Star Formation Studies #Classification scheme #Pattern recognition (psychology) #Robustness (evolution) #Satellite #Scheme (mathematics) #astro-ph
paper · pdf · doi:10.1086/420967
published as Astrophys.J.Suppl. 152 (2004) 201 · 26 pages, To appear in ApJS after July 2004
arxiv created 2004/03/10 · openalex publication_date 2004/05/19 · arxiv updated 2009/12/01 · openalex created_date 2016/06/24 · openalex updated_date 2026/08/05
An artificial neural network (ANN) scheme has been employed that uses a supervised back-propagation algorithm to classify 2000 bright sources from the Calgary database of Infrared Astronomical Satellite ( IRAS ) spectra in the region 8-23 μm. The database has been classified into 17 predefined classes based on the spectral morphology. We have been able to classify over 80% of the sources correctly in the first instance. The speed and robustness of the scheme will allow us to classify the whole of the Low Resolution Spectrometer database, containing more than 50,000 sources, in the near future.