Identifying correlations between LIGO's astronomical range and auxiliary sensors using lasso regression. (19th October 2018)

• Record Type: Journal Article
• Authors: Walker, Marissa; Agnew, Alfonso F; Bidler, Jeffrey; Lundgren, Andrew; Macedo, Alexandra; Macleod, Duncan; Massinger, T J; Patane, Oliver; Smith, Joshua R
• Is Part Of: Classical and quantum gravity. Volume 35:Number 22(2018:Nov.)
• Publication Date: 2018-10-19
• Subjects: gravitational wave detectors -- machine learning -- regression; Quantum gravity -- Periodicals; Gravitation -- Periodicals; Relativity (Physics) -- Periodicals; Space and time -- Periodicals; Periodicals; 521.1
• DOI: 10.1088/1361-6382/aae593
• Languages: English
• ISSNs: 0264-9381

Abstract

Abstract: The range to which the Laser Interferometer Gravitational-Wave Observatory (LIGO) can observe astrophysical systems varies over time, limited by noise in the instruments and their environments. Identifying and removing the sources of noise that limit LIGO's range enables higher signal-to-noise observations and increases the number of observations. The LIGO observatories are continuously monitored by hundreds of thousands of auxiliary channels that may contain information about these noise sources. This paper describes an algorithm that uses linear regression, namely lasso (least absolute shrinkage and selection operator) regression, to analyze all of these channels and identify a small subset of them that can be used to reconstruct variations in LIGO's astrophysical range. Exemplary results of the application of this method to three different periods of LIGO Livingston data are presented, along with computational performance and current limitations.