Abstract
Bayesian Blocks is a technique for detecting and characterizing signals in noisy time series. This time-domain method establishes a representation with some features of wavelet expansions, but at the same time relaxing some of their restrictions. With Bayesian Blocks all details of the representation are flexible and determined by the data through optimization of a piecewise constant model. As with wavelets, Bayesian Blocks can effect denoising without explicit smoothing and the concomitant loss of information through degraded resolution.
Get full access to this article
View all access options for this article.