Time series clustering based on forecast densities

被引：59

作者：

Alonso, A. M.

Berrendero, J. R.

Hernandez, A.

Justel, A. ^{[1
]}

机构：

[1] Univ Autonoma Madrid, Dept Matemat, Madrid, Spain

[2] Univ Carlos III Madrid, Dept Estadist, Madrid, Spain

[3] Univ Exeter, Dept Math Sci, Exeter, Devon, England

来源：

COMPUTATIONAL STATISTICS & DATA ANALYSIS | 2006年 / 51卷 / 02期

关键词：

bootstrap; cluster analysis; L-2-distance; Kyoto protocol; nonparametric density estimation; prediction;

D O I：

10.1016/j.csda.2006.04.035

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

A new clustering method for time series is proposed, based on the full probability density of the forecasts. First, a resampling method combined with a nonparametric kernel estimator provides estimates of the forecast densities. A measure of discrepancy is then defined between these estimates and the resulting dissimilarity matrix is used to carry out the required cluster analysis. Applications of this method to both simulated and real life data sets are discussed. (c) 2006 Elsevier B.V. All rights reserved.

引用

页码：762 / 776

页数：15

共 26 条

[1] On sieve bootstrap prediction intervals [J].