The sensitivity of searches for astrophysical transients in data from the LIGO is generally limited by the presence of transient, non-Gaussian noise artifacts, which occur at a high-enough rate such that accidental coincidence across multiple detecto
rs is non-negligible. Furthermore, non-Gaussian noise artifacts typically dominate over the background contributed from stationary noise. These glitches can easily be confused for transient gravitational-wave signals, and their robust identification and removal will help any search for astrophysical gravitational-waves. We apply Machine Learning Algorithms (MLAs) to the problem, using data from auxiliary channels within the LIGO detectors that monitor degrees of freedom unaffected by astrophysical signals. The number of auxiliary-channel parameters describing these disturbances may also be extremely large; an area where MLAs are particularly well-suited. We demonstrate the feasibility and applicability of three very different MLAs: Artificial Neural Networks, Support Vector Machines, and Random Forests. These classifiers identify and remove a substantial fraction of the glitches present in two very different data sets: four weeks of LIGOs fourth science run and one week of LIGOs sixth science run. We observe that all three algorithms agree on which events are glitches to within 10% for the sixth science run data, and support this by showing that the different optimization criteria used by each classifier generate the same decision surface, based on a likelihood-ratio statistic. Furthermore, we find that all classifiers obtain similar limiting performance, suggesting that most of the useful information currently contained in the auxiliary channel parameters we extract is already being used.
There is a broad class of astrophysical sources that produce detectable, transient, gravitational waves. Some searches for transient gravitational waves are tailored to known features of these sources. Other searches make few assumptions about the so
urces. Typically events are observable with multiple search techniques. This work describes how to combine the results of searches that are not independent, treating each search as a classifier for a given event. This will be shown to improve the overall sensitivity to gravitational-wave events while directly addressing the problem of consistent interpretation of multiple trials.
We describe a general approach to detection of transient gravitational-wave signals in the presence of non-Gaussian background noise. We prove that under quite general conditions, the ratio of the likelihood of observed data to contain a signal to th
e likelihood of it being a noise fluctuation provides optimal ranking for the candidate events found in an experiment. The likelihood-ratio ranking allows us to combine different kinds of data into a single analysis. We apply the general framework to the problem of unifying the results of independent experiments and the problem of accounting for non-Gaussian artifacts in the searches for gravitational waves from compact binary coalescence in LIGO data. We show analytically and confirm through simulations that in both cases the likelihood ratio statistic results in an improved analysis.
We report the results of the first search for gravitational waves from compact binary coalescence using data from the LIGO and Virgo detectors. Five months of data were collected during the concurrent S5 (LIGO) and VSR1 (Virgo) science runs. The sear
ch focused on signals from binary mergers with a total mass between 2 and 35 Msun. No gravitational waves are identified. The cumulative 90%-confidence upper limits on the rate of compact binary coalescence are calculated for non-spinning binary neutron stars, black hole-neutron star systems, and binary black holes to be 8.7x10^-3, 2.2x10^-3 and 4.4x10^-4 yr^-1 L_10^-1 respectively, where L_10 is 10^10 times the blue solar luminosity. These upper limits are compared with astrophysical expectations.
