We present a Bayesian parameter-estimation pipeline to measure the properties of inspiralling stellar-mass black hole binaries with LISA. Our strategy (i) is based on the coherent analysis of the three noise-orthogonal LISA data streams, (ii) employs
accurate and computationally efficient post-Newtonian waveforms accounting for both spin-precession and orbital eccentricity, and (iii) relies on a nested sampling algorithm for the computation of model evidences and posterior probability density functions of the full 17 parameters describing a binary. We demonstrate the performance of this approach by analyzing the LISA Data Challenge (LDC-1) dataset, consisting of 66 quasi-circular, spin-aligned binaries with signal-to-noise ratios ranging from 3 to 14 and times to merger ranging from 3000 to 2 years. We recover 22 binaries with signal-to-noise ratio higher than 8. Their chirp masses are typically measured to better than $0.02 M_odot$ at $90%$ confidence, while the sky-location accuracy ranges from 1 to 100 square degrees. The mass ratio and the spin parameters can only be constrained for sources that merge during the mission lifetime. In addition, we report on the successful recovery of an eccentric, spin-precessing source at signal-to-noise ratio 15 for which we can measure an eccentricity of $3times 10^{-3}$.
The population of Milky Way satellite galaxies is of great interest for cosmology, fundamental physics, and astrophysics. They represent the faint end of the galaxy luminosity function, are the most dark-matter dominated objects in the local Universe
, and contain the oldest and most metal-poor stellar populations. Recent surveys have revealed around 60 satellites, but this could represent less than half of the total. Characterization of these systems remains a challenge due to their low luminosity. We consider the gravitational wave observatory LISA as a potential tool for studying these satellites through observations of their short-period double white dwarf populations. LISA will observe the entire sky without selection effects due to dust extinction, complementing optical surveys, and could potentially discover massive satellites hidden behind the disk of the galaxy.
The label switching problem arises in the Bayesian analysis of models containing multiple indistinguishable parameters with arbitrary ordering. Any permutation of these parameters is equivalent, therefore models with many such parameters have extreme
ly multi-modal posterior distributions. It is difficult to sample efficiently from such posteriors. This paper discusses a solution to this problem which involves carefully mapping the input parameter space to a high dimensional hypertriangle. It is demonstrated that this solution is efficient even for large numbers of parameters and can be easily applied alongside any stochastic sampling algorithm. This method is illustrated using two example problems from the field of gravitational wave astronomy.
