Abstract

This paper describes the Context Tree Switching technique, a modification of Context Tree
\nWeighting for the prediction of binary, stationary, n-Markov sources. By modifying Context
\nTree Weighting’s recursive weighting scheme, it is possible to mix over a strictly larger class of
\nmodels without increasing the asymptotic time or space complexity of the original algorithm.
\nWe prove that this generalization preserves the desirable theoretical properties of Context Tree
\nWeighting on stationary n-Markov sources, and show empirically that this new technique leads
\nto consistent improvements over Context Tree Weighting as measured on the Calgary Corpus.

Keywords

WeightingTree (set theory)Context (archaeology)Computer scienceBinary treeMarkov chainMarkov processGeneralizationClass (philosophy)AlgorithmMathematicsTheoretical computer scienceArtificial intelligenceMachine learningStatisticsCombinatorics

Affiliated Institutions

Related Publications

Publication Info

Year
2012
Type
article
Volume
47
Pages
327-336
Citations
40
Access
Closed

External Links

View on DOI.org

Social Impact

Altmetric
PlumX Metrics

Social media, news, blog, policy document mentions

Citation Metrics

40
OpenAlex

Cite This

Joel Veness, Kee Siong Ng, Marcus Hütter et al. (2012). Context Tree Switching. , 47 , 327-336. https://doi.org/10.1109/dcc.2012.39

Identifiers

DOI
10.1109/dcc.2012.39