The BayesWave analysis pipeline in the era of gravitational wave observations. (arXiv:2011.09494v2 [gr-qc] UPDATED)
<a href="http://arxiv.org/find/gr-qc/1/au:+Cornish_N/0/1/0/all/0/1">Neil J. Cornish</a>, <a href="http://arxiv.org/find/gr-qc/1/au:+Littenberg_T/0/1/0/all/0/1">Tyson B. Littenberg</a>, <a href="http://arxiv.org/find/gr-qc/1/au:+Becsy_B/0/1/0/all/0/1">Bence Bécsy</a>, <a href="http://arxiv.org/find/gr-qc/1/au:+Chatziioannou_K/0/1/0/all/0/1">Katerina Chatziioannou</a>, <a href="http://arxiv.org/find/gr-qc/1/au:+Clark_J/0/1/0/all/0/1">James A. Clark</a>, <a href="http://arxiv.org/find/gr-qc/1/au:+Ghonge_S/0/1/0/all/0/1">Sudarshan Ghonge</a>, <a href="http://arxiv.org/find/gr-qc/1/au:+Millhouse_M/0/1/0/all/0/1">Margaret Millhouse</a>
We describe updates and improvements to the BayesWave gravitational wave
transient analysis pipeline, and provide examples of how the algorithm is used
to analyze data from ground-based gravitational wave detectors. BayesWave
models gravitational wave signals in a morphology-independent manner through a
sum of frame functions, such as Morlet-Gabor wavelets or chirplets. BayesWave
models the instrument noise using a combination of a parametrized Gaussian
noise component and non-stationary and non-Gaussian noise transients. Both the
signal model and noise model employ trans-dimensional sampling, with the
complexity of the model adapting to the requirements of the data. The
flexibility of the algorithm makes it suitable for a variety of analyses,
including reconstructing generic unmodeled signals; cross checks against
modeled analyses for compact binaries; as well as separating coherent signals
from incoherent instrumental noise transients (glitches). The BayesWave model
has been extended to account for gravitational wave signals with generic
polarization content and the simultaneous presence of signals and glitches in
the data. We describe updates in the BayesWave prior distributions, sampling
proposals, and burn-in stage that provide significantly improved sampling
efficiency. We present standard review checks indicating the robustness and
convergence of the BayesWave trans-dimensional sampler.
We describe updates and improvements to the BayesWave gravitational wave
transient analysis pipeline, and provide examples of how the algorithm is used
to analyze data from ground-based gravitational wave detectors. BayesWave
models gravitational wave signals in a morphology-independent manner through a
sum of frame functions, such as Morlet-Gabor wavelets or chirplets. BayesWave
models the instrument noise using a combination of a parametrized Gaussian
noise component and non-stationary and non-Gaussian noise transients. Both the
signal model and noise model employ trans-dimensional sampling, with the
complexity of the model adapting to the requirements of the data. The
flexibility of the algorithm makes it suitable for a variety of analyses,
including reconstructing generic unmodeled signals; cross checks against
modeled analyses for compact binaries; as well as separating coherent signals
from incoherent instrumental noise transients (glitches). The BayesWave model
has been extended to account for gravitational wave signals with generic
polarization content and the simultaneous presence of signals and glitches in
the data. We describe updates in the BayesWave prior distributions, sampling
proposals, and burn-in stage that provide significantly improved sampling
efficiency. We present standard review checks indicating the robustness and
convergence of the BayesWave trans-dimensional sampler.
http://arxiv.org/icons/sfx.gif
