Dimension reduction for stationary multivariate time series data

Fayed Alshammri; Jiazhu Pan

Dimension reduction for stationary multivariate time series data

Fayed Alshammri, Jiazhu Pan

Mathematics And Statistics

Research output: Contribution to conference › Poster

7 Downloads (Pure)

Abstract

Chang et al. (2016) extended PCA by finding a linear transformation of the original variables such that the transformed series is segmented into uncorrelated subseries with lower dimensions. This method is called TS-PCA. In our current research, we will extend TS-PCA by reducing the dimension of the transformed subseries further by applying GDPCA by Pena and Yohai (2016) to the results from TS-PCA, and possibly reach a further dimension reduction. Hence, the proposed method is a combination of TS-PCA and GDPCA.

Original language	English
Number of pages	1
Publication status	Published - 19 May 2017
Event	The Education, Research, Humanities, and Statistics International Conference - Washington DC, United States Duration: 19 May 2017 → 19 May 2019

Conference

Conference	The Education, Research, Humanities, and Statistics International Conference
Country/Territory	United States
City	Washington DC
Period	19/05/17 → 19/05/19

Keywords

multivariate time series
TS-PCA
dimension reduction

Access to Document

Alshammri-Pan-ERHS2017-Dimension-reduction-for-stationary-multivariate-time-series-dataFinal published version, 1 MB

Best Talk Prize
Alshammri, Fayed Awdah M (Recipient), 2017
Prize: Prize (including medals and awards)

Cite this

@conference{dbbbbf54b8ba4584a853ecdac41727e4,

title = "Dimension reduction for stationary multivariate time series data",

abstract = "Chang et al. (2016) extended PCA by finding a linear transformation of the original variables such that the transformed series is segmented into uncorrelated subseries with lower dimensions. This method is called TS-PCA. In our current research, we will extend TS-PCA by reducing the dimension of the transformed subseries further by applying GDPCA by Pena and Yohai (2016) to the results from TS-PCA, and possibly reach a further dimension reduction. Hence, the proposed method is a combination of TS-PCA and GDPCA.",

keywords = "multivariate time series, TS-PCA, dimension reduction",

author = "Fayed Alshammri and Jiazhu Pan",

year = "2017",

month = may,

day = "19",

language = "English",

note = "The Education, Research, Humanities, and Statistics International Conference ; Conference date: 19-05-2017 Through 19-05-2019",

}

TY - CONF

T1 - Dimension reduction for stationary multivariate time series data

AU - Alshammri, Fayed

AU - Pan, Jiazhu

PY - 2017/5/19

Y1 - 2017/5/19

N2 - Chang et al. (2016) extended PCA by finding a linear transformation of the original variables such that the transformed series is segmented into uncorrelated subseries with lower dimensions. This method is called TS-PCA. In our current research, we will extend TS-PCA by reducing the dimension of the transformed subseries further by applying GDPCA by Pena and Yohai (2016) to the results from TS-PCA, and possibly reach a further dimension reduction. Hence, the proposed method is a combination of TS-PCA and GDPCA.

AB - Chang et al. (2016) extended PCA by finding a linear transformation of the original variables such that the transformed series is segmented into uncorrelated subseries with lower dimensions. This method is called TS-PCA. In our current research, we will extend TS-PCA by reducing the dimension of the transformed subseries further by applying GDPCA by Pena and Yohai (2016) to the results from TS-PCA, and possibly reach a further dimension reduction. Hence, the proposed method is a combination of TS-PCA and GDPCA.

KW - multivariate time series

KW - TS-PCA

KW - dimension reduction

M3 - Poster

T2 - The Education, Research, Humanities, and Statistics International Conference

Y2 - 19 May 2017 through 19 May 2019

ER -

Dimension reduction for stationary multivariate time series data

Abstract

Conference

Keywords

Access to Document

Fingerprint

Prizes

Best Talk Prize

Cite this