Title :
Adaptive Context Tree Weighting
Author :
O´Neill, Alexander ; Hutter, Marcus ; Shao, Wen ; Sunehag, Peter
Abstract :
We describe an adaptive context tree weighting (ACTW) algorithm, as an extension to the standard context tree weighting (CTW) algorithm. Unlike the standard CTW algorithm, which weights all observations equally regardless of the depth, ACTW gives increasing weight to more recent observations, aiming to improve performance in cases where the input sequence is from a non-stationary distribution. Data compression results show ACTW variants improving over CTW on merged files from standard compression benchmark tests while never being significantly worse on any individual file.
Keywords :
benchmark testing; data compression; encoding; trees (mathematics); adaptive context tree weighting; data compression; input sequence; nonstationary distribution; standard compression benchmark tests; Algorithm design and analysis; Bayesian methods; Context; Data compression; Encoding; History; Prediction algorithms;
Conference_Titel :
Data Compression Conference (DCC), 2012
Conference_Location :
Snowbird, UT
Print_ISBN :
978-1-4673-0715-4
DOI :
10.1109/DCC.2012.38
