Fast segmentation algorithms for long hydrometeorological time series

被引:31
作者
Aksoy, Hafzullah [1 ]
Gedikli, Abdullah [2 ]
Unal, N. Erdem [1 ]
Kehagias, Athanasios [3 ]
机构
[1] Istanbul Tech Univ, Dept Civil Engn, Hydraul Div, TR-34469 Istanbul, Turkey
[2] Istanbul Tech Univ, Dept Civil Engn, Div Appl Mech, TR-34469 Istanbul, Turkey
[3] Aristotle Univ Thessaloniki, Sch Engn, GR-54124 Thessaloniki, Greece
关键词
time series; segmentation; change point; dynamic programming; branch-and-bound approach;
D O I
10.1002/hyp.7064
中图分类号
TV21 [水资源调查与水利规划];
学科分类号
081501 ;
摘要
A time series with natural or artificially created inhomogeneities can be segmented into parts with different statistical characteristics. In this study, three algorithms are presented for time series segmentation; the first is based on dynamic programming and the second and the third-the latter being an improved version of the former-are based on the branch-and-bound approach. The algorithms divide the time series into segments using the first order statistical moment (average). Tested on real world time series of several hundred or even over a thousand terms the algorithms performs segmentations satisfactorily and fast. (c) Copyright 2008 John Wiley & Sons, Ltd.
引用
收藏
页码:4600 / 4608
页数:9
相关论文
共 24 条
[11]  
Hipel KW., 1994, TIME SERIES MODELLIN
[12]   The segmentation procedure as a tool for discrete modeling of hydrometeorological regimes [J].
Hubert, P .
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT, 2000, 14 (4-5) :297-304
[13]   A dynamic programming segmentation procedure for hydrological and environmental time series [J].
Kehagias, A ;
Nidelkou, E ;
Petridis, V .
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT, 2006, 20 (1-2) :77-94
[14]   A hidden Markov model segmentation procedure for hydrological and environmental time series [J].
Kehagias, A .
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT, 2004, 18 (02) :117-130
[15]  
Kehagias A, 2007, STOCH ENV RES RISK A, V21, P451, DOI 10.1007/s00477-006-0092-z
[16]   Time series segmentation with shifting means hidden Markov models [J].
Kehagias, Ath. ;
Fortin, V. .
NONLINEAR PROCESSES IN GEOPHYSICS, 2006, 13 (03) :339-352
[17]   Northern hemisphere temperatures during the past millennium: Inferences, uncertainties, and limitations [J].
Mann, ME ;
Bradley, RS ;
Hughes, MK .
GEOPHYSICAL RESEARCH LETTERS, 1999, 26 (06) :759-762
[18]  
Mann MichaelE., 2000, EARTH INTERACT, V4, P1, DOI DOI 10.1175/1087-3562(2000)004ANDLT
[19]  
0001:GTPIPCANDGT
[20]  
2.3.CO