Presentation #130.02 in the session Star Associations.
We will present first results from our high-dimensional correlation search among parameters related to star cluster formation and evolution, as obtained by employing state-of-the-art machine learning techniques. Understanding the physics underlying the star formation cycle is an inherently multi-phase and multi-scale problem requiring data from across the electromagnetic spectrum. The PHANGS (Physics at High Angular Resolution in Nearby GalaxieS) program is combining observations with ALMA, VLT/MUSE (PHANGS-ALMA and PHANGS-MUSE, PI: E. Schinnerer) and HST (PHANGS-HST, PI: J. C. Lee) to chart the connections between molecular clouds, HII regions, and young stars. Owing to the resolving power of HST, the PHANGS-HST project is obtaining a census of star clusters across 38 nearby spiral galaxies. We are producing catalogs including the physical (e.g., their ages and masses) and quantitative morphological (e.g., M20, multiple concentration index) properties of tens of thousands of star clusters. Combined with their local (galactic radius, placement on a spiral arm or a bar, and similar) and global (such as the SFH of the host galaxy) environmental properties, and observations from ALMA and MUSE, we have constructed a high-dimensional data set of substantial volume on which to systematically analyze the parameters that influence the formation and evolution of these basic units of star formation. We will present results employing both supervised and unsupervised learning algorithms analyzing this wealth of data, the insights they provide for the formation and evolution of star clusters, and compare to well-known empirical scaling relationships that have driven the field for many decades.