Title :
On delayed prediction of individual sequences
Author :
Weinberger, Marcelo J. ; Ordentlich, Erik
Author_Institution :
Hewlett Packard Labs., Palo Alto, CA, USA
Abstract :
We investigate a prediction scenario in which the predictor is forced to make a decision a number of steps in advance, with incomplete information. For finite action and observation spaces, it is shown that the strategy that minimizes the worst-case regret with respect to the Bayes envelope is obtained through sub-sampling of the sequence of observations. The result extends to the case of logarithmic loss.
Keywords :
Bayes methods; information theory; minimax techniques; prediction theory; sequences; Bayes envelope; delayed prediction; finite action; individual sequences; logarithmic loss; minimax strategy; observation spaces; sub-sampling; Convergence; Delay effects; Delta modulation; Laboratories; Minimax techniques; Performance loss; Probability distribution; Sampling methods;
Conference_Titel :
Information Theory, 2002. Proceedings. 2002 IEEE International Symposium on
Print_ISBN :
0-7803-7501-7
DOI :
10.1109/ISIT.2002.1023420
