http://arxiv.org/abs/1908.08677
The LAMOST survey has provided 9 million spectra in its Data Release 5 (DR5) at R~1800. Extracting precise stellar labels is crucial for such a large sample. In this paper, we report the implementation of the Stellar LAbel Machine (SLAM), which is a data-driven model based on Support Vector Regression (SVR), a robust non-linear regression technique. Thanks to the capability to model a highly non-linear problem with SVR, SLAM generally can derive stellar labels over a wide range of spectral types. This gives it unique capability compared to other popular data-driven models. To illustrate this capability, we test the performance of SLAM for stars ranging from Teff ~4000 to ~8000K trained by LAMOST spectra with stellar labels from the LAMOST pipeline. At g-band signal-to-noise ratio (SNRg) higher than 100, the random uncertainties of Teff, log g and [Fe/H] are 50 K, 0.09 dex, and 0.07 dex, respectively. We then set up another SLAM model trained by APOGEE and LAMOST common stars to demonstrate its capability of dealing with high dimensional problems. The spectra are from LAMOST DR5 and the stellar labels of the training dataset are from APOGEE DR15, including Teff , log g, [M/H], [a/M], [C/M], and [N/M]. The cross-validated scatters at SNRg ~ 100 are 49 K, 0.10 dex, 0.037 dex, 0.026 dex, 0.058 dex, and 0.106 dex for these parameters, respectively. This performance is at the same level as other up-to-date data-driven models. We therefore conclude that SLAM is ready for deriving multi-dimensional stellar labels in high precision over a larger range of spectral types. As a byproduct, we also provide the latest catalog of ~ 1 million LAMOST DR5 K giant stars with SLAM-predicted stellar labels in this work.
B. Zhang, C. Liu and L. Deng
Mon, 26 Aug 19
1/55
Comments: 18 pages, 11 figures, submitted to ApJS