2011/10/20 by Scott F. Daniel, Andrew J. Connolly, Jeff Schneider +2
Computer Science · Physics and Astronomy · #Artificial intelligence #Astronomical spectroscopy #Astronomy #Astronomy and Astrophysical Research #Astrophysics #Blind Source Separation Techniques #Computer science #Curse of dimensionality #Dimensionality reduction #Embedding #Pattern recognition (psychology) #Physics #Spectral line #Stars #Stellar classification #Stellar, planetary, and galactic studies #astro-ph.SR
paper · pdf · doi:10.1088/0004-6256/142/6/203
published as The Astronomical Journal, 142 (2011) 203 · 15 pages, 13 figures; accepted for publication in The Astronomical Journal
arxiv created 2011/10/20 · openalex publication_date 2011/11/15 · arxiv updated 2011/11/16 · openalex created_date 2016/06/24 · openalex updated_date 2026/08/05
We investigate the use of dimensionality reduction techniques for the classification of stellar spectra selected from the Sloan Digital Sky Survey. Using local linear embedding (LLE), a technique that preserves the local (and possibly nonlinear) structure within high-dimensional data sets, we show that the majority of stellar spectra can be represented as a one-dimensional sequence within a three-dimensional space. The position along this sequence is highly correlated with spectral temperature. Deviations from this "stellar locus" are indicative of spectra with strong emission lines (including misclassified galaxies) or broad absorption lines (e.g., carbon stars). Based on this analysis, we propose a hierarchical classification scheme using LLE that progressively identifies and classifies stellar spectra in a manner that requires no feature extraction and that can reproduce the classic MK classifications to an accuracy of one type.