Unsupervised Machine Learning Classification of Stellar Populations in Open Clusters Using Gaia DR3 Photometry and Astrometry

This study applies four unsupervised machine learning algorithms - k-means, DBSCAN, Gaussian Mixture Models, and Self-Organizing Maps - to Gaia DR3 photometry and astrometry for 5,409 confirmed members across four open clusters (Pleiades, Praesepe, NGC 2516, and Blanco 1) spanning a seven-fold age range of 100-700 Myr. Starting from 714,540 raw sources, a five-stage quality filtering pipeline retained confirmed cluster members via parallax, proper motion, and astrometric quality cuts. All four algorithms were applied to a standardized five-dimensional feature space comprising BP-RP color, absolute G magnitude, parallax, and both proper motion components. The pipeline recovered an age-dependent main sequence turnoff progression across all four clusters from photometry alone, spontaneously detected giant and supergiant candidates in NGC 2516 and white dwarf candidates in multiple clusters without targeted searches, and revealed genuine density discontinuities in five-dimensional feature space through SOM U-matrix analysis. Cross-method validation via PARSEC isochrone overlays at literature ages confirmed pipeline accuracy and membership validity. These results demonstrate that unsupervised machine learning applied to public Gaia photometry and astrometry, without spectroscopic data or manual intervention, can recover physically meaningful stellar evolutionary structure, providing a scalable foundation for systematic stellar population analysis across the thousands of open clusters in the Gaia catalog.