Bayesian model selection applied to linear regressions with weighted data

This paper uses the Bayesian model-selection to find the linear regression that is most supported by the data. It uses a hierarchical model to improve and to extend to weighted data from the results of a previous investigation. As an application example, it shows how the results can be used to investigate the consistency of measurements carried out to determine the central second-moment of the angular power-spectrum of a laser beam.