In press as of 2/3/2013:
STUDIES IN ASTRONOMICAL TIME SERIES ANALYSIS.
VI. BAYESIAN BLOCK REPRESENTATIONS
Jeffrey D. Scargle1, Jay P. Norris2, Brad Jackson3, and James Chiang
The Astrophysical Journal, 764:2 (26pp), 2013, p. 167
doi:10.1088/0004-637X/763/1/1
“The line is similar to a length of time, and as the points are the beginning and end of the line, so the instants are the endpoints of any given extension of time.” Leonardo da Vinci, Codex Arundel, folio 190v., c. 1500
The paper and MatLab code to implement the algorithm and reproduce all of the figures in the paper (that is, implementing "Reproducible Research" -- Donoho, D., Maleki, A., Rahman, I., Shahram, M., & Stodden, V. Fifteen Years of Reproducible Research in Computational Harmonic Analysis, 2009, CSE, 11, 8) are also posted at Studies in Astronomical Time Series Analysis: VI. Bayesian Block Representations in the astro-ph archive.
Updates and corrections to the code, new applications and references, etc. will be posted on the blog from time to time. Two bugs (affecting the display of event data block representations and the "empty block" features) have been identified and fixes will be posted soon.
Jake Vanderplas' related blog Dynamic Programming in Python: Bayesian Blocks nicely described the algorithmic approach of dynamic programming, with examples in the context of histograms.
Jake Vanderplas' related blog Dynamic Programming in Python: Bayesian Blocks nicely described the algorithmic approach of dynamic programming, with examples in the context of histograms.
