Refined M-type star catalog from LAMOST DR10

In this work, we refined the LAMOST DR10 M-type star catalog through a two-stage process. First, we purified the catalog using techniques including deep learning and color-magnitude diagrams to remove 22,496 non-M spectra, corrected 2078 dwarf/giant classifications, and updated 12,900 radial velocities.

This resulted in a cleaner catalog containing 870,518 M-type spectra (820,493 dwarfs, 50,025 giants). Second, applying a label transfer strategy using values from APOGEE DR16 for parameter prediction with a tenfold cross-validated convolutional neural network ensemble architecture, we predicted Teff, logg, [M/H], and [{alpha}/M] separately for M dwarfs (dMs) and giants (gMs). The average internal errors for dMs/gMs are, respectively, Teff 30/17K, logg 0.07/0.07dex, [M/H] 0.07/0.05dex, and [{alpha}/M] 0.02/0.02dex.