The Javalambre Photometric Local Universe Survey (J-PLUS) has obtained precise photometry in twelve specially designed filters for large numbers of Galactic stars. Deriving their precise stellar atmospheric parameters and individual elemental abundances is crucial for studies of Galactic structure, and the assembly history and chemical evolution of our Galaxy. We have estimated stellar parameters (effective temperature, Teff , surface gravity, log g, and metallicity, [Fe/H]), [α/Fe], and four elemental abundances ([C/Fe], [N/Fe], [Mg/Fe], and [Ca/Fe]) by a set of cost-sensitive neural networks (CSNet) from recalibrated J-PLUS DR1.

The results are summarized here and fully detailed in the paper:

The stellar parameters (effective temperature, Teff, surface gravity, log g, and metallicity, [Fe/H]), [α/Fe], and four elemental abundances ([C/Fe], [N/Fe], [Mg/Fe], and [Ca/Fe]) from J-PLUS DR1 are estimated by a set of cost-sensitive neural networks (CSNet), as shown in Fig.1. The dataset used for training and testing CSNet is constructed by stars in common between J-PLUS DR1 (Cenarro et al. 2019), Gaia DR2 (Gaia Collaboration et al. 2018), and LAMOST DR5 (Luo et al. 2015; Xiang et al. 2019), where the first two data sets provide input stellar colors and the last one provides stellar labels (Teff , log g, [Fe/H], [C/Fe], [N/Fe], [Mg/Fe], [Ca/Fe] and [α/Fe]). The catalog contains 4,387,568 stars (MAGABDUALOBJ_CLASS_STAR ≥ 0.6) by cross-matching J-PLUS DR1 with Gaia DR2. We recommend stellar label estimates with FLAGS=0 and X_FLAG=0 for about two million stars. Furthermore, only stellar label estimates located in the same range with the training set (see Table 1) are considered to be reliable, as we are cautious to extrapolate, a well-known limitation of ANN-based approaches. To avoid large label uncertainties caused by photometric errors, particularly for elemental abundances, labels for stars with magnitude errors in the 12 J-PLUS filters more than 0.1 mag should be used with caution.

Table 1. Adopted constraints on the datasets for training and testing CSNet.

Data download

How to cite

If you have used this data in your paper, please cite the following paper:

• Yang et al. 2021