The problem of backing out a star-formation history (SFH) from an optical spectrum is an ill-posed one, hampered by degeneracies like that between stellar age, stellar metallicity, and foreground dust attenuation. Therefore, it is hard to estimate what mass is contained in a stellar population. Meanwhile, a galaxy's mass is one of the best indicators of its evolutionary state, so it is important to make a good measurement!

We have devised a training dataset of 40,000 SFHs, generated theoretical spectra, for each, and found the best representation of this ensemble of spectra as 6 mutually-orthogonal eigenspectra. This basis set eases the model-data comparison, and allows us to estimate stellar mass-to-light ratio for integral-field spectroscopic (IFS) data from the SDSS-IV/MaNGA survey.

In Pace et al. 2019a ("Paper I"), we detail the training data, the PCA method, and the quality-assurance metrics computed. In Pace et al. 2019b ("Paper II"), we present several applications of these data, including a comparison to dynamical masses found through high-resolution, vertical kinematics of face-on disk galaxies, aperture-corrections, and a discussion of the consequences of spatial binning.

In SDSS Data Release 16, we released a Value-Added Catalog of resolved stellar mass maps and integrated, aperture-corrected stellar masses for ~4700 MaNGA galaxies.