Exploratory analysis of functional data via clustering and optimal segmentation

被引:31
作者
Hebrail, Georges [1 ]
Hugueney, Bernard [2 ]
Lechevallier, Yves [3 ]
Rossi, Fabrice [1 ]
机构
[1] Telecom ParisTech, BILab, CNRS, UMR 5141,LTCI, F-75013 Paris, France
[2] Univ Paris 09, LAMSADE, F-75016 Paris, France
[3] INRIA, Projet AxIS, F-78153 Le Chesnay, France
关键词
Functional data; Multiple time series; Exploratory analysis; Clustering; Segmentation; Dynamic programming; ALGORITHMS; SELECTION; CURVES;
D O I
10.1016/j.neucom.2009.11.022
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We propose in this paper an exploratory analysis algorithm for functional data. The method partitions a set of functions into K clusters and represents each cluster by a simple prototype (e.g., piecewise constant). The total number of segments in the prototypes, P, is chosen by the user and optimally distributed among the clusters via two dynamic programming algorithms. The practical relevance of the method is shown on two real world datasets. (C) 2010 Elsevier B.V. All rights reserved.
引用
收藏
页码:1125 / 1141
页数:17
相关论文
共 35 条
  • [1] Unsupervised curve clustering using B-splines
    Abraham, C
    Cornillon, PA
    Matzner-Lober, E
    Molinari, N
    [J]. SCANDINAVIAN JOURNAL OF STATISTICS, 2003, 30 (03) : 581 - 595
  • [2] [Anonymous], LECT NOTES COMPUTER
  • [3] [Anonymous], 1979, SIAMR, DOI DOI 10.1137/1021044
  • [4] [Anonymous], 1997, FUNCTIONAL DATA ANAL
  • [5] [Anonymous], 2001, SPRINGER SERIES INFO, DOI DOI 10.1007/978-3-642-56927-2
  • [6] AUGER IE, 1989, B MATH BIOL, V51, P39, DOI 10.1007/BF02458835
  • [7] ON THE APPROXIMATION OF CURVES BY LINE SEGMENTS USING DYNAMIC PROGRAMMING
    BELLMAN, R
    [J]. COMMUNICATIONS OF THE ACM, 1961, 4 (06) : 284 - 284
  • [8] CHAMROUKHI F, 2009, P 15 EUR S ART NEUR, P503
  • [9] ENTROPY-BASED ALGORITHMS FOR BEST BASIS SELECTION
    COIFMAN, RR
    WICKERHAUSER, MV
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 1992, 38 (02) : 713 - 718
  • [10] Cottrell M, 1998, J FORECASTING, V17, P429, DOI 10.1002/(SICI)1099-131X(1998090)17:5/6<429::AID-FOR706>3.0.CO