The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
L'un des avantages des méthodes du noyau est qu'elles peuvent traiter différents types d'objets, pas nécessairement des données vectorielles avec un nombre fixe d'attributs. Dans cet article, nous développons des noyaux pour les données de séries chronologiques en utilisant des distances de déformation temporelle dynamique (DTW). Étant donné que les distances DTW sont des pseudo-distances qui ne satisfont pas à l’inégalité triangulaire, une matrice de noyau basée sur celles-ci n’est pas semi-définie positive, en général. Nous utilisons la programmation semi-définie (SDP) pour garantir le caractère défini positif d'une matrice de noyau. Nous présentons l'intégration préservant le voisinage (NPE), une formulation SDP pour obtenir une matrice de noyau qui préserve au mieux la géométrie locale des données de séries chronologiques. Nous présentons également une extension hors échantillon (OSE) pour NPE. Nous utilisons deux applications, la classification de séries chronologiques et l'intégration de séries chronologiques pour la recherche de similarité, pour valider notre approche.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Hiroyuki NARITA, Yasumasa SAWAMURA, Akira HAYASHI, "DTW-Distance Based Kernel for Time Series Data" in IEICE TRANSACTIONS on Information,
vol. E92-D, no. 1, pp. 51-58, January 2009, doi: 10.1587/transinf.E92.D.51.
Abstract: One of the advantages of the kernel methods is that they can deal with various kinds of objects, not necessarily vectorial data with a fixed number of attributes. In this paper, we develop kernels for time series data using dynamic time warping (DTW) distances. Since DTW distances are pseudo distances that do not satisfy the triangle inequality, a kernel matrix based on them is not positive semidefinite, in general. We use semidefinite programming (SDP) to guarantee the positive definiteness of a kernel matrix. We present neighborhood preserving embedding (NPE), an SDP formulation to obtain a kernel matrix that best preserves the local geometry of time series data. We also present an out-of-sample extension (OSE) for NPE. We use two applications, time series classification and time series embedding for similarity search, to validate our approach.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E92.D.51/_p
Copier
@ARTICLE{e92-d_1_51,
author={Hiroyuki NARITA, Yasumasa SAWAMURA, Akira HAYASHI, },
journal={IEICE TRANSACTIONS on Information},
title={DTW-Distance Based Kernel for Time Series Data},
year={2009},
volume={E92-D},
number={1},
pages={51-58},
abstract={One of the advantages of the kernel methods is that they can deal with various kinds of objects, not necessarily vectorial data with a fixed number of attributes. In this paper, we develop kernels for time series data using dynamic time warping (DTW) distances. Since DTW distances are pseudo distances that do not satisfy the triangle inequality, a kernel matrix based on them is not positive semidefinite, in general. We use semidefinite programming (SDP) to guarantee the positive definiteness of a kernel matrix. We present neighborhood preserving embedding (NPE), an SDP formulation to obtain a kernel matrix that best preserves the local geometry of time series data. We also present an out-of-sample extension (OSE) for NPE. We use two applications, time series classification and time series embedding for similarity search, to validate our approach.},
keywords={},
doi={10.1587/transinf.E92.D.51},
ISSN={1745-1361},
month={January},}
Copier
TY - JOUR
TI - DTW-Distance Based Kernel for Time Series Data
T2 - IEICE TRANSACTIONS on Information
SP - 51
EP - 58
AU - Hiroyuki NARITA
AU - Yasumasa SAWAMURA
AU - Akira HAYASHI
PY - 2009
DO - 10.1587/transinf.E92.D.51
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E92-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2009
AB - One of the advantages of the kernel methods is that they can deal with various kinds of objects, not necessarily vectorial data with a fixed number of attributes. In this paper, we develop kernels for time series data using dynamic time warping (DTW) distances. Since DTW distances are pseudo distances that do not satisfy the triangle inequality, a kernel matrix based on them is not positive semidefinite, in general. We use semidefinite programming (SDP) to guarantee the positive definiteness of a kernel matrix. We present neighborhood preserving embedding (NPE), an SDP formulation to obtain a kernel matrix that best preserves the local geometry of time series data. We also present an out-of-sample extension (OSE) for NPE. We use two applications, time series classification and time series embedding for similarity search, to validate our approach.
ER -