Tracking the Optimal Sequence of Predictive Strategies

V. V. V’yugin, V. G. Trunov

Research output: Contribution to journalArticlepeer-review

Abstract

Abstract: Within the prediction (decision making) theory with online experts, an adaptive algorithm is proposed that aggregates the decisions of expert strategies and incurs losses that do not exceed (up to a certain value, called a regret) the losses of the best combination of experts distributed over the prediction interval. The algorithm develops the Mixing Past Posteriors method and the AdaHedge algorithm of exponential weighting of expert decisions using an adaptive learning parameter. An estimate of the regret of the proposed algorithm is obtained. The approach proposed does not make assumptions about the nature of the data source and the limits of experts’ losses. The results of numerical experiments on mixing expert solutions using the proposed algorithm under conditions of high volatility of experts’ losses are given.

Original languageEnglish
Pages (from-to)1491-1501
Number of pages11
JournalJournal of Communications Technology and Electronics
Volume63
Issue number12
DOIs
Publication statusPublished - 1 Dec 2018
Externally publishedYes

Keywords

  • adaptive learning parameter
  • mixing schemes for posterior expert distributions
  • online loss distribution algorithms
  • predictions using expert strategies

Fingerprint

Dive into the research topics of 'Tracking the Optimal Sequence of Predictive Strategies'. Together they form a unique fingerprint.

Cite this