Bayesian Analysis Techniques for Gravitational Wave Data

Open Access
Breysse, Patrick Conrad
Area of Honors:
Bachelor of Science
Document Type:
Thesis Supervisors:
  • Lee S Finn, Thesis Supervisor
  • Richard Wallace Robinett, Honors Advisor
  • Gravitational Waves
  • LIGO
  • Data Analysis
  • Bayesian Statistics
Data from gravitational wave detectors suffer from a very low signal to noise ratio, which makes it difficult to find signals using conventional statistical techniques. Fitting a data set with conventional techniques will often fit measurement noise in addition to signal, leading to incorrect results. We explore a solution to this problem using Bayesian model comparison techniques. A given data set is first fit with each of a variety of models using Bayes' law, then the models are ranked based on their evidence, which is a measure of the probability that the model accurately represents the source of the data. The model fitting stage has been shown to be effective on a sample problem of radioactive decay. It was then applied to looking for the signal of black hole ringdown. The evidence ranking was also tested on the ringdown problem, but it has not yet been proven to be effective.