Context tree estimation for not necessarily finite memory processes, via BIC and MDL.
Imre CsiszárZsolt TalataPublished in: IEEE Trans. Inf. Theory (2006)
Keyphrases
- model selection
- maximum likelihood
- penalized likelihood
- contextual information
- tree structure
- parameter estimation
- context aware
- minimum description length
- accurate estimation
- context sensitive
- estimation algorithm
- sample size
- binary tree
- memory space
- closed form
- data sets
- index structure
- mixture model
- machine learning