Estimating Stellar Parameters from LAMOST Low-resolution Spectra

Abstract

The Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) has acquired tens of millions of low-resolution spectra of stars. This paper investigated the parameter estimation problem for these spectra. To this end, we proposed a deep learning model StarGRU network (StarGRUNet). This network was further applied to estimate the stellar atmospheric physical parameters and 13 elemental abundances from LAMOST low-resolution spectra. On the spectra with signal-to-noise ratios greater than or equal to $5$, the estimation precisions are $94$ K and $0.16$ dex on $T_\texttt{eff}$ and $\log \ g$ respectively, $0.07$ dex to $0.10$ dex on [C/H], [Mg/H], [Al/H], [Si/H], [Ca/H], [Ni/H] and [Fe/H], and $0.10$ dex to $0.16$ dex on [O/H], [S/H], [K/H], [Ti/H] and [Mn/H], and $0.18$ dex and $0.22$ dex on [N/H] and [Cr/H] respectively. The model shows advantages over available models and high consistency with high-resolution surveys. We released the estimated catalog computed from about 8.21 million low-resolution spectra in LAMOST DR8, code, trained model, and experimental data for astronomical science exploration and data processing algorithm research respectively.