Title :
Compression-Based Methods for Nonparametric Prediction and Estimation of Some Characteristics of Time Series
Author_Institution :
Inst. of Comput. Technol., Russian Acad. of Sci., Novosibirsk, Russia
Abstract :
We address the problem of online prediction for time series. We show that any universal code (or a universal data compressor) can be used as a basis for constructing asymptotically optimal methods for this problem for a certain class of stationary and ergodic processes.
Keywords :
data compression; estimation theory; nonparametric statistics; time series; asymptotically optimal methods; compression-based methods; ergodic processes; nonparametric estimation; nonparametric prediction; online prediction; time series; universal code; universal data compressor; Data compression; H infinity control; Informatics; Information theory; Predictive models; Random processes; Source coding; Statistics; Telecommunication computing; Testing; Density estimation; prediction of random processes; source coding; stationary ergodic source; universal coding;
Journal_Title :
Information Theory, IEEE Transactions on
DOI :
10.1109/TIT.2009.2025546
