Combining Expert Advice Efficiently. / Koolen, Wouter; de Rooij, Steven.

2008.

Research output: Working paper

Published

Standard

Combining Expert Advice Efficiently. / Koolen, Wouter; de Rooij, Steven.

2008.

Research output: Working paper

Harvard

APA

Vancouver

Author

BibTeX

@techreport{1c198d02246c4f3fab6708e503b4cd6f,
title = "Combining Expert Advice Efficiently",
abstract = "We show how models for prediction with expert advice can be defined concisely and clearly using hidden Markov models (HMMs); standard HMM algorithms can then be used to efficiently calculate, among other things, how the expert predictions should be weighted according to the model. We cast many existing models as HMMs and recover the best known running times in each case. We also describe two new models: the switch distribution, which was recently developed to improve Bayesian/Minimum Description Length model selection, and a new generalisation of the fixed share algorithm based on run-length coding. We give loss bounds for all models and shed new light on their relationships.",
keywords = "cs.LG, cs.DS, cs.IT, math.IT, G.3",
author = "Wouter Koolen and {de Rooij}, Steven",
note = "50 pages",
year = "2008",
month = "2",
day = "14",
language = "Undefined/Unknown",
type = "WorkingPaper",

}

RIS

TY - UNPB

T1 - Combining Expert Advice Efficiently

AU - Koolen, Wouter

AU - de Rooij, Steven

N1 - 50 pages

PY - 2008/2/14

Y1 - 2008/2/14

N2 - We show how models for prediction with expert advice can be defined concisely and clearly using hidden Markov models (HMMs); standard HMM algorithms can then be used to efficiently calculate, among other things, how the expert predictions should be weighted according to the model. We cast many existing models as HMMs and recover the best known running times in each case. We also describe two new models: the switch distribution, which was recently developed to improve Bayesian/Minimum Description Length model selection, and a new generalisation of the fixed share algorithm based on run-length coding. We give loss bounds for all models and shed new light on their relationships.

AB - We show how models for prediction with expert advice can be defined concisely and clearly using hidden Markov models (HMMs); standard HMM algorithms can then be used to efficiently calculate, among other things, how the expert predictions should be weighted according to the model. We cast many existing models as HMMs and recover the best known running times in each case. We also describe two new models: the switch distribution, which was recently developed to improve Bayesian/Minimum Description Length model selection, and a new generalisation of the fixed share algorithm based on run-length coding. We give loss bounds for all models and shed new light on their relationships.

KW - cs.LG

KW - cs.DS

KW - cs.IT

KW - math.IT

KW - G.3

M3 - Working paper

BT - Combining Expert Advice Efficiently

ER -