A probabilistic framework for cosmological inference of peculiar velocities

We present a Bayesian hierarchical framework for a principled data analysis pipeline of peculiar velocity surveys, which makes explicit the inference problem of constraining cosmological parameters from redshift-independent distance indicators. We demonstrate our method for a Fundamental Plane (FP)-based survey. The essence of our approach is to work closely with observables (e.g. angular size, surface brightness, redshift, etc.), through which we bypass the use of summary statistics by working with the probability distributions. The hierarchical approach improves upon the usual analysis in several ways. In particular, it allows a consistent analysis without having to make prior assumptions about cosmology during the calibration phase. Moreover, calibration uncertainties are correctly accounted for in parameter estimation. Results are presented for a new, fully analytic posterior marginalized over all latent variables, which we expect to allow for more principled analyses in upcoming surveys. A maximum a posteriori estimator is also given for peculiar velocities derived from FP data.