Инд. авторы:	Ryabko B.Y.
Заглавие:	Compression-based Methods for Nonparametric Density Estimation, On-Line Prediction, Regression and Classification for Time Series.
Библ. ссылка:	Ryabko B.Y. Compression-based Methods for Nonparametric Density Estimation, On-Line Prediction, Regression and Classification for Time Series. // 2008 IEEE Information Theory Workshop: IEEE Information Theory Workshop (MAY 05-08, 2008). - 2008. - Porto. - P.271-275. - ISBN: 978-1-4244-2269-2.
Внешние системы:	DOI: 10.1109/ITW.2008.4578667; РИНЦ: 15053389; SCOPUS: 2-s2.0-51849157344; WoS: 000259299000061;
Реферат:	eng: We address the problem of nonparametric estimation of characteristics for stationary and ergodic time series. We consider finite-alphabet time series and the real-valued ones and the following problems: estimation of the (limiting) probability P(u(0) ... u(s)) for every s and each sequence u(0) ... u(s) of letters from the process alphabet (or estimation of the density p(x(0), ..., x(s)) for real-valued time series), so-called on-line prediction, where the conditional probability P(x(t+1)/x(1)x(2) ... x(t)) (or the conditional density p(x(t+1)/x(1)x(2) ... x(t))) should be estimated (in the case where x(1)x(2) ... x(t) is known), regression and classification (or so-called problems with side information). We show that any universal code (or a universal data compressor) can be used as a basis for constructing asymptotically optimal methods for the above problems.
Ключевые слова:	Time-series; On-line prediction; Nonparametric density estimation; Non-parametric estimations; Ergodic; Time series analysis; Ketones; Information theory; Food processing; Estimation; Cybernetics; Maximum likelihood estimation;
Издано:	Porto: , 2008
Физ. характеристика:	с.271-275
Конференция:	Название: 2008 IEEE Information Theory Workshop Город: Porto Страна: Portugal Даты проведения: 2008-05-05 - 2008-05-08