EmissionGaiaXP: Data and Models for Hα Emission Detection in Gaia XP Spectra

Datasets and trained machine-learning models accompanying the paper 'From Active Stars to Black Holes II: Discovery of Hα Emission in Ultra-Low Resolution Gaia XP Spectra with Machine Learning' (Rodriguez et al., PAI 2026). Includes the labeled training set (75,789 Gaia DR3 sources cross-matched with SDSS and LAMOST and labeled by Hα equivalent width), processed coefficient arrays, trained classifiers (logistic regression, MLP, 1D CNN, XGBoost), large unlabeled Gaia XP application samples, model predictions, and X-ray (eROSITA) cross-match candidate lists. Code to reproduce the analysis is on GitHub: https://github.com/acrodrig98/EmissionGaiaXP