Identification of Ap Candidates from LAMOST DR12 Based on Deep Learning

Ap stars are a subclass of early-type chemically peculiar stars characterized by abnormal surface elemental abundances and strong, global magnetic fields. These features make them valuable tracers for studying stellar magnetic field origins and evolution. In recent years, the identification of Ap stars has become a research hotspot. We present a novel method using low-resolution LAMOST DR12 spectra. We developed ECACNet, a deep neural network incorporating the ECA mechanism to enhance sensitivity to key spectral features of magnetic Ap stars. On a balanced dataset covering seven stellar classes, ECACNet achieved 96.80% precision and 94.05% recall. Evaluated on seven independent validation sets, it maintained robust performance (96.13% precision, 92.03% recall), significantly outperforming conventional methods. Applying the model to LAMOST DR12 B- to F-type spectra identified 8,130 high-confidence spectra. Visual inspection confirmed 4,826 unique Ap candidates via the characteristic 5200 Å flux depression and metallic lines. Cross-matching with eight catalogs revealed 4,032 known and 794 new candidates. MKCLASS refined the new candidates’ spectral classifications. Statistical analyses of atmospheric parameters (Teff, log g, and [Fe/H]) for the new candidates showed distributions consistent with known Ap stars, further validating our method’s reliability.