Abstract
In recent years, dynamic time warping (DTW) has begun to become the most widely used technique for comparison of time series data where extensive a priori knowledge is not available. However, it is often expected a multivariate comparison method to consider the correlation between the variables as this correlation carries the real information in many cases. Thus, principal component analysis (PCA) based similarity measures, such as PCA similarity factor (SPCA), are used in many industrial applications. In this paper, we present a novel algorithm called correlation based dynamic time warping (CBDTW) which combines DTW and PCA based similarity measures. To preserve correlation, multivariate time series are segmented and the local dissimilarity function of DTW originated from SPCA. The segments are obtained by bottom-up segmentation using special, PCA related costs. Our novel technique qualified on two databases, the database of signature verification competition 2004 and the commonly used AUSLAN dataset. We show that CBDTW outperforms the standard SPCA and the most commonly used, Euclidean distance based multivariate DTW in case of datasets with complex correlation structure.
Original language | English |
---|---|
Pages (from-to) | 12814-12823 |
Number of pages | 10 |
Journal | Expert Systems with Applications |
Volume | 39 |
Issue number | 17 |
DOIs | |
Publication status | Published - Dec 1 2012 |
Fingerprint
Keywords
- Dynamic time warping
- Multivariate time series
- Principal component analysis
- Segmentation
- Similarity
ASJC Scopus subject areas
- Artificial Intelligence
- Computer Science Applications
- Engineering(all)
Cite this
Correlation based dynamic time warping of multivariate time series. / Bankó, Zoltán; Abonyi, J.
In: Expert Systems with Applications, Vol. 39, No. 17, 01.12.2012, p. 12814-12823.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Correlation based dynamic time warping of multivariate time series
AU - Bankó, Zoltán
AU - Abonyi, J.
PY - 2012/12/1
Y1 - 2012/12/1
N2 - In recent years, dynamic time warping (DTW) has begun to become the most widely used technique for comparison of time series data where extensive a priori knowledge is not available. However, it is often expected a multivariate comparison method to consider the correlation between the variables as this correlation carries the real information in many cases. Thus, principal component analysis (PCA) based similarity measures, such as PCA similarity factor (SPCA), are used in many industrial applications. In this paper, we present a novel algorithm called correlation based dynamic time warping (CBDTW) which combines DTW and PCA based similarity measures. To preserve correlation, multivariate time series are segmented and the local dissimilarity function of DTW originated from SPCA. The segments are obtained by bottom-up segmentation using special, PCA related costs. Our novel technique qualified on two databases, the database of signature verification competition 2004 and the commonly used AUSLAN dataset. We show that CBDTW outperforms the standard SPCA and the most commonly used, Euclidean distance based multivariate DTW in case of datasets with complex correlation structure.
AB - In recent years, dynamic time warping (DTW) has begun to become the most widely used technique for comparison of time series data where extensive a priori knowledge is not available. However, it is often expected a multivariate comparison method to consider the correlation between the variables as this correlation carries the real information in many cases. Thus, principal component analysis (PCA) based similarity measures, such as PCA similarity factor (SPCA), are used in many industrial applications. In this paper, we present a novel algorithm called correlation based dynamic time warping (CBDTW) which combines DTW and PCA based similarity measures. To preserve correlation, multivariate time series are segmented and the local dissimilarity function of DTW originated from SPCA. The segments are obtained by bottom-up segmentation using special, PCA related costs. Our novel technique qualified on two databases, the database of signature verification competition 2004 and the commonly used AUSLAN dataset. We show that CBDTW outperforms the standard SPCA and the most commonly used, Euclidean distance based multivariate DTW in case of datasets with complex correlation structure.
KW - Dynamic time warping
KW - Multivariate time series
KW - Principal component analysis
KW - Segmentation
KW - Similarity
UR - http://www.scopus.com/inward/record.url?scp=84865005617&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84865005617&partnerID=8YFLogxK
U2 - 10.1016/j.eswa.2012.05.012
DO - 10.1016/j.eswa.2012.05.012
M3 - Article
AN - SCOPUS:84865005617
VL - 39
SP - 12814
EP - 12823
JO - Expert Systems with Applications
JF - Expert Systems with Applications
SN - 0957-4174
IS - 17
ER -