The 35th annual conference on Astronomical Data Analysis Software & Systems 2025 in Görlitz, Germany

Classifying and summarizing extensive datasets from diverse sky surveys is essential for advancing astronomical research. Integrating data from 4XMM-DR13 (X-ray), SDSS DR18 (optical), and CatWISE (IR) surveys, we constructed the XMM-WISE-SDSS sample. Cross-matching with SDSS/LAMOST spectral classifications provided a training set of stars, galaxies, quasars, and young stellar objects (YSOs). We classified the full sample using CatBoost and Self-Paced Ensemble (SPE) machine learning. The SPE classifier excelled in YSO identification, detecting 1,102 YSO candidates—including 258 known YSOs. Verification using LAMOST spectra and SIMBAD/VizieR databases confirmed 412 new YSO candidates. These discoveries substantially expand the known YSO sample, enabling deeper studies of star formation and evolution. A comprehensive classification catalog for the XMM-WISE-SDSS sample is provided.