Large-scale typicality of Markov sample paths and consistency of MDL order estimators

IEEE Transactions on Information Theory

Volumn 48, Issue 6, 2002, Pages 1616-1628

  Csiszár, Imre ^a,b  

^a IEEE   (Hungary)

^b INSTITUTE OF EXPERIMENTAL MEDICINE   (Hungary)

Author keywords

Conditional typicality; Empirical distribution; Iterated logarithm; Markov chain; Minimum description length (MDL); Normalized maximum likelihood (NML); Order estimation

Indexed keywords

CONDITIONAL TYPICALITY;

ITERATIVE METHODS; MARKOV PROCESSES; MAXIMUM LIKELIHOOD ESTIMATION; PROBABILITY; THEOREM PROVING;

INFORMATION THEORY;

EID: 0036611605     PISSN: 00189448     EISSN: None     Source Type: Journal    
DOI: 10.1109/TIT.2002.1003842     Document Type: Article

References (12)

1. The minimum description length principle in coding and modeling
2. Variable length Markov chains
3. The consistency of the BIC Markov order estimator
4. Deviations from uniformity in random strings
5. The performance of universal encoding
6. Coding of discrete sources with unknown statistics
7. A universal finite memory source
8. The context tree weighting method: Basic properties