Abstract:Observation of gravitational waves from inspiralling binary black holes has offered a unique opportunity to study the physical parameters of the component black holes. To infer these parameters, Bayesian methods are employed in conjunction with general relativistic waveform models that describe the source's inspiral, merger, and ringdown. The results depend not only on the accuracy of the waveform models but also on the underlying fiducial prior distribution used for the analysis. In particular, when the pre-merger phase of the signal is barely observable within the detectors' bandwidth, as is currently the case with intermediate-mass black hole binary signals in ground-based gravitational wave detectors, different prior assumptions can lead to different interpretations. In this study, we utilise the gravitational-wave inference library, $\texttt{Parallel Bilby}$, to evaluate the impact of mass prior choices on the parameter estimation of intermediate-mass black hole binary signals. While previous studies focused primarily on analysing event data, we offer a broader, more controlled study by using simulations. Our findings suggest that the posteriors in total mass, mass ratio and luminosity distance are contingent on the assumed mass prior distribution used during the inference process. This is especially true when the signal lacks sufficient pre-merger information and/or has inadequate power in the higher-order radiation multipoles. In conclusion, our study underscores the importance of thoroughly investigating similarly heavy events in current detector sensitivity using a diverse choice of priors. Absent such an approach, adopting a flat prior on the binary's redshifted total mass and mass ratio emerges as a reasonable choice, preventing biases in the detector-frame mass posteriors.

An introduction to Bayesian inference in gravitational-wave astronomy: Parameter estimation, model selection, and hierarchical models

Bilby: A User-friendly Bayesian Inference Library for Gravitational-wave Astronomy

Bayesian parameter estimation for targeted anisotropic gravitational-wave background

Growing Pains: Understanding the Impact of Likelihood Uncertainty on Hierarchical Bayesian Inference for Gravitational-Wave Astronomy

Inferring Binary Properties from Gravitational Wave Signals

Practical Guidance for Bayesian Inference in Astronomy

Incorporating model accuracy into gravitational-wave Bayesian inference

Gravitational-wave parameter estimation with relative binning: Inclusion of higher-order modes and precession, and applications to lensing and third-generation detectors

Matching Bayesian and frequentist coverage probabilities when using an approximate data covariance matrix

Improving the scalability of Gaussian-process error marginalization in gravitational-wave inference

Impact of Bayesian Priors on the Inferred Masses of Quasi-Circular Intermediate-Mass Black Hole Binaries

Calibrating approximate Bayesian credible intervals of gravitational-wave parameters

Detecting the Undetected: Overcoming Biases in Gravitational-Wave Population Studies

Bayesian inference in physics

Accelerated Bayesian parameter estimation and model selection for gravitational waves with normalizing flows

Accounting for Selection Effects in Supernova Cosmology with Simulation-Based Inference and Hierarchical Bayesian Modelling

Bayesian Computation in Astronomy: Novel methods for parallel and gradient-free inference

Pulsar Timing Arrays require hierarchical models

Massively parallel Bayesian inference for transient gravitational-wave astronomy

BayesLine: Bayesian Inference for Spectral Estimation of Gravitational Wave Detector Noise