Studies in Astronomical Time Series Analysis. VI. Optimal Segmentation: Blocks, Triggers, and Histograms

This paper addresses the problem of detecting and characterizing local variability in time series. Since such data are always corrupted by observational errors, the goal is to find statistically any significant variations and ignore the inevitable random noise fluctuations. We present a simple nonparametric modeling technique and an algorithm implementing it—an improved and generalized version of Bayesian Blocks [Scargle 1998]—that finds the optimal partitioning of the observation interval. The structure of the algorithm allows it to be used in either a real-time, trigger mode, or a retrospective mode. The necessary maximum likelihood or marginal posterior functions to measure model fitness are presented for points, binned counts, and measurements at arbitrary times with a known error distribution. The same algorithm can also be used to compute histograms.