Mass and Age Determination of the LAMOST Data with Different Machine-learning Methods

Abstract

We present a catalog of 948,216 stars with mass labels and a catalog of 163,105 red clump (RC) stars with mass and age labels simultaneously. The training data set is crossmatched from the Large Sky Area Multi-Object Fiber Spectroscopic Telescope DR5, and high-resolution asteroseismology data, mass, and age are predicted by the random forest (RF) method or a convex-hull algorithm. The stellar parameters with a high correlation with mass and age are extracted and the test data set shows that the median relative error of the prediction model for the mass of the large sample is 3%, and for the mass and age of RC stars is 4% and 7%. We also compare the predicted age of RC stars with recent works and find that the final uncertainty of the RC sample could reach 18% for age and 9% for mass; meanwhile, the final precision of the mass for the large sample with different types of stars could reach 13% without considering systematics. All of this implies that this method could be widely used in the future. Moreover, we explore the performance of different machine-learning methods for our sample, including Bayesian linear regression and the gradient-boosting decision tree (GBDT), multilayer perceptron, multiple linear regression, RF, and support vector regression methods. Finally, we find that the performance of a nonlinear model is generally better than that of a linear model, and the GBDT and RF methods are relatively better.

Age Determination of LAMOST Red Giant Branch Stars Based on the Gradient Boosting Decision Tree Method

“…Similar accuracy was obtained earlier by Mackereth et al (2019) using a Bayesian convolution neural network and APOGEE DR14 data trained on APOGEE-Kepler data (Pinsonneault et al 2018). Li et al (2022) used random forest methods and convex-hull algorithms to predict RGB and RCG masses based on scikitlearn (Pedregosa et al 2011), and hence, RCG ages. The median relative error is 13% for a large sample of K giant masses and 9% and 18% for RC stellar masses and ages, respectively.…”

Evidence for Population-dependent Vertical Motions and the Long-lived Nonsteady Lopsided Milky Way Warp

“…Since we only focus on the range of R = 8-14 kpc, the error in the distance is acceptable. The details of the RC star sample can be found in the work of Ting et al (2018), which includes a catalog of 175,202 RC stars, with an uncertainty of the distance within 10%.This catalog is also widely used to explore the stellar mass and age (Li et al 2022). Our MSTO samples and distances are based on the work of Xiang et al (2017aXiang et al ( , 2017b, which includes a catalog of 670,714 MSTO stars, and the error in the distance is estimated to be 10%-30%.…”

Kinematic-chemical Analysis and Time Tagging for the Diagonal Ridge Structure of the Galactic Outer Disk with LAMOST Red-giant Branch Stars

“…The radial velocity uncertainty is 5 km s −1 estimated by using the LAMOST stellar parameter pipeline of Peking University (Xiang et al 2017), the error of mass and age determined by the KPCA is 10% and 30%, respectively. Metallicity error is 0.1 dex, [α/Fe] error is about 0.05 dex (which are also wildly used in our previous work; Li et al 2022), and the distance uncertainty is about 15%. The proper-motion information of the sample comes from the Gaia DR2 catalog, and the uncertainties of the proper motion is 0.06 mas yr −1 (for G < 15 mag), 0.2 mas yr −1 (for G = 17 mag) and 1.2 mas yr −1 (for G = 20 mag; Gaia Collaboration et al 2018).…”