Time series classification using local distance-based features in multi-modal fusion networks

Brian Kenji Iwana, Seiichi Uchida

Research output: Contribution to journalArticle

Abstract

We propose the use of a novel feature, called local distance features, for time series classification. The local distance features are extracted using Dynamic Time Warping (DTW) and classified using Convolutional Neural Networks (CNN). DTW is classically as a robust distance measure for distance-based time series recognition methods. However, by using DTW strictly as a global distance measure, information about the matching is discarded. We show that this information can further be used as supplementary input information in temporal CNNs. This is done by using both the raw data and the features extracted from DTW in multi-modal fusion CNNs. Furthermore, we explore the effects of different prototype selection methods, prototype numbers, and data fusion schemes induce on the accuracy. We perform experiments on a wide range of time series datasets including three Unipen handwriting datasets, four UCI Machine Learning Repository datasets, and 85 UCR Time Series Classification Archive datasets.

Original languageEnglish
Article number107024
JournalPattern Recognition
Volume97
DOIs
Publication statusPublished - Jan 1 2020

Fingerprint

Data fusion
Convolution
Clustering algorithms
Time series
Neural networks
Learning systems
Experiments

All Science Journal Classification (ASJC) codes

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

Time series classification using local distance-based features in multi-modal fusion networks. / Kenji Iwana, Brian; Uchida, Seiichi.

In: Pattern Recognition, Vol. 97, 107024, 01.01.2020.

Research output: Contribution to journalArticle

@article{8bb8a9ebde224a079ebf56aaff8605c1,
title = "Time series classification using local distance-based features in multi-modal fusion networks",
abstract = "We propose the use of a novel feature, called local distance features, for time series classification. The local distance features are extracted using Dynamic Time Warping (DTW) and classified using Convolutional Neural Networks (CNN). DTW is classically as a robust distance measure for distance-based time series recognition methods. However, by using DTW strictly as a global distance measure, information about the matching is discarded. We show that this information can further be used as supplementary input information in temporal CNNs. This is done by using both the raw data and the features extracted from DTW in multi-modal fusion CNNs. Furthermore, we explore the effects of different prototype selection methods, prototype numbers, and data fusion schemes induce on the accuracy. We perform experiments on a wide range of time series datasets including three Unipen handwriting datasets, four UCI Machine Learning Repository datasets, and 85 UCR Time Series Classification Archive datasets.",
author = "{Kenji Iwana}, Brian and Seiichi Uchida",
year = "2020",
month = "1",
day = "1",
doi = "10.1016/j.patcog.2019.107024",
language = "English",
volume = "97",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Time series classification using local distance-based features in multi-modal fusion networks

AU - Kenji Iwana, Brian

AU - Uchida, Seiichi

PY - 2020/1/1

Y1 - 2020/1/1

N2 - We propose the use of a novel feature, called local distance features, for time series classification. The local distance features are extracted using Dynamic Time Warping (DTW) and classified using Convolutional Neural Networks (CNN). DTW is classically as a robust distance measure for distance-based time series recognition methods. However, by using DTW strictly as a global distance measure, information about the matching is discarded. We show that this information can further be used as supplementary input information in temporal CNNs. This is done by using both the raw data and the features extracted from DTW in multi-modal fusion CNNs. Furthermore, we explore the effects of different prototype selection methods, prototype numbers, and data fusion schemes induce on the accuracy. We perform experiments on a wide range of time series datasets including three Unipen handwriting datasets, four UCI Machine Learning Repository datasets, and 85 UCR Time Series Classification Archive datasets.

AB - We propose the use of a novel feature, called local distance features, for time series classification. The local distance features are extracted using Dynamic Time Warping (DTW) and classified using Convolutional Neural Networks (CNN). DTW is classically as a robust distance measure for distance-based time series recognition methods. However, by using DTW strictly as a global distance measure, information about the matching is discarded. We show that this information can further be used as supplementary input information in temporal CNNs. This is done by using both the raw data and the features extracted from DTW in multi-modal fusion CNNs. Furthermore, we explore the effects of different prototype selection methods, prototype numbers, and data fusion schemes induce on the accuracy. We perform experiments on a wide range of time series datasets including three Unipen handwriting datasets, four UCI Machine Learning Repository datasets, and 85 UCR Time Series Classification Archive datasets.

UR - http://www.scopus.com/inward/record.url?scp=85071504668&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071504668&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2019.107024

DO - 10.1016/j.patcog.2019.107024

M3 - Article

VL - 97

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

M1 - 107024

ER -