Determination of Hubble constant from Megamaser Cosmology Project using Profile Likelihood

TL;DR

This paper demonstrates that using profile likelihood provides an independent and consistent frequentist estimate of the Hubble constant from Megamaser data, validating its effectiveness in cosmological parameter inference.

Contribution

It introduces the application of profile likelihood to cosmology, offering an alternative to Bayesian methods for handling nuisance parameters in Hubble constant estimation.

Findings

Frequentist estimate of H0: 73.5^{+3.0}_{-2.9} km/sec/Mpc

Estimate agrees with Bayesian result within 0.2 sigma

Profile likelihood effectively manages nuisance parameters in cosmology

Abstract

The Megamaser Cosmology Project inferred a value for the Hubble constant given by $H_0=73.9 \pm 3.0 $ km/sec/Mpc. This value was obtained using Bayesian inference by marginalizing over six nuisance parameters, corresponding to the velocities of the megamaser galaxy systems. We obtain an independent estimate of the Hubble constant with the same data using frequentist inference. For this purpose, we use profile likelihood to dispense with the aforementioned nuisance parameters. The frequentist estimate of the Hubble constant is given by $H_0=73.5^{+3.0}_{-2.9}$ km/sec/Mpc and agrees with the Bayesian estimate to within $0.2\sigma$, and both approaches also produce consistent confidence/credible intervals. Therefore, this analysis provides a proof-of-principle application of profile likelihood in dealing with nuisance parameters in cosmology, which is complementary to Bayesian analysis.