The connection between continuum emission and emission lines in galaxy spectra
Keywords:emission lines, galaxy spectra, machine learning
We present a machine learning procedure to investigate the connection between emission lines and the spectral continuum in the spectrum of galaxies that is agnostic with respect to the source of line excitation: star formation or nuclear activity. We know a priori that there is some connection between a galaxy continuum and its pattern of emission lines. We explore this connection with a deep learning model, establishing a mapping between the galactic continuum and the equivalent emission line widths for the sample discussed by Werle et al. (2019) of SDSS galaxies with GALEX photometry. We showed that we are able to estimate equivalent widths with good accuracy and, consequently, we were able to partially reproduce the BPT diagram. This suggests that there is indeed a relationship between the source of ionization responsible for the emission lines and the stellar populations that produce the galactic continuum.