Supplementary data for "Revealing Ambiguity and Misclassification in Gaia DR3 Cepheid and RR Lyrae Variable Stars: A Triangulated Machine Learning Approach"

This deposit contains supplementary data for "Revealing Ambiguity and Substructure in Gaia DR3 Cepheid and RR Lyrae Classification: A Triangulated Machine Learning Approach" (Wu, 2026). RNAAS. flagged_catalog_of_ambiguous_cepheid_variables.csv — Catalog of 71 Cepheid stars flagged as ambiguous or misclassified, identified via disagreement between supervised ML predictions, unsupervised HDBSCAN clustering, and literature/catalog cross-check. Columns include Gaia DR3 ID, RA, DEC, Gaia Class, New ML Pred, #ML Disagree, Clustering Status, SIMBAD Class. ML_model_performance_eval.csv — Per-class precision, recall, F1 scores, and support across six supervised ML models (RFC, KNN, DTC, GNB, SVM, NN) evaluated on Gaia DR3 Cepheid and RR Lyrae subtypes (DCEP, T2CEP, ACEP, RRab, RRc, RRd) via 3-fold nested cross-validation. ml_confusion_matrix.csv — Confusion matrices for all six supervised ML models (Random Forest, SVM, Decision Tree, KNN, Naive Bayes, Neural Network), stacked in long format by model and true_class, with columns as predicted_class.