Minimax optimal Bayes mixtures for memoryless sources over large alphabets

MaRDI QIDQ4617620zbMATH OpenFDO

Authors Elias Jääsaari, Janne Leppä-aho, Tomi Silander, Teemu Roos

Publication date 6 February 2019

Full work available at URL http://proceedings.mlr.press/v83/jaasaari18a/jaasaari18a.pdf

zbMATH Keywords

Bayes mixturelarge alphabetminimax regretnormalized maximum likelihoodonline learninguniversal codinguniversal prediction

Mathematics Subject Classification ID

Statistical aspects of information-theoretic topics (62B10) Bayesian problems; characterization of Bayes procedures (62C10) Minimax procedures in statistical decision theory (62C20)

Recommendations

• Asymptotic minimax regret for data compression, gambling, and prediction
• Sequential universal modeling for non-binary sequences with constrained distributions
• Achievability of asymptotic minimax regret by horizon-dependent and horizon-independent strategies
• Minimax redundancy for the class of memoryless sources
• Arbitrary source models and Bayesian codebooks in rate-distortion theory

Cited in

(6)

• Greedy structure learning from data that contain systematic missing values
• Multi-task transfer learning for Bayesian network structures
• On the MDL principle for i.i.d. sources with large alphabets
• Asymptotic minimax regret for data compression, gambling, and prediction
• Minimum description length revisited
• Achievability of asymptotic minimax regret by horizon-dependent and horizon-independent strategies

This page was built for publication: Minimax optimal Bayes mixtures for memoryless sources over large alphabets

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4617620)