Using Kepler data and machine learning to help improve constraints on global parameters of red giants observed with K2 and TESS
James Kuszlewicz (University of Birmingham), Guy R. Davies (University of Birmingham)

The wealth of red giant stars observed with Kepler enables us to extract a great deal of information on their stellar structure and dynamics. By using machine learning techniques we can use results from the nominal Kepler mission to constrain the stellar properties of both lower signal-to-noise and shorter length datasets, as it is under these situations where the determination of the background parameters (and more importantly, global seismic parameters) can be much more difficult. We employ Bayesian statistics and a supervised machine learning scheme to derive a set of priors from background fits to high signal-to-noise APOKASC red giants, which are used as a training set. With this newly extracted information the resultant priors were then applied in similar fits to help constrain K2 C1 and C3 data, paving the way to further explore models of the Milky Way using asteroseismology.