Title :
Informational confidence bounds for self-normalized averages and applications
Author :
Garivier, Aurelien
Author_Institution :
Inst. de Math. de Toulouse, Univ. Paul Sabatier, Toulouse, France
Abstract :
We present deviation bounds for self-normalized averages and applications to estimation with a random number of observations. The results rely on a peeling argument in exponential martingale techniques that represents an alternative to the method of mixture. The motivating examples of bandit problems and context tree estimation are detailed.
Keywords :
estimation theory; random processes; stochastic processes; bandit problem; context tree estimation; exponential martingale technique; informational confidence deviation bound; peeling argument; random number estimation; self-normalized average; Context; Data models; Equations; Estimation; Markov processes; Mathematical model;
Conference_Titel :
Information Theory Workshop (ITW), 2013 IEEE
Conference_Location :
Sevilla
Print_ISBN :
978-1-4799-1321-3
DOI :
10.1109/ITW.2013.6691311