TY - JOUR
T1 - A Review of Distributed Algorithms for Principal Component Analysis
AU - Wu, Sissi Xiaoxiao
AU - Wai, Hoi To
AU - Li, Lin
AU - Scaglione, Anna
N1 - Funding Information:
Manuscript received February 24, 2018; revised May 27, 2018; accepted June 4, 2018. Date of current version August 2, 2018. This work was supported in part by the U.S. Air Force under Contracts FA8721-05-C-0002 and/or FA8702-15-D-0001; by the National Natural Science Foundation of China under Grant 61701315; by Shenzhen Technology R&D Fund JCYJ20170817101149906 and JCYJ20170302145906843; by Shenzhen University Launch Fund 2018018; and by the U.S. National Science Foundation under Grants EAGER CCF 1553746, NSF CCF-BSF 1714672, and BSF 2016660. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the U.S. Government. (Corresponding author: Sissi Xiaoxiao Wu.) S. X. Wu is with the Department of Communication and Information Engineering, Shenzhen University, China (e-mail: xxwu.eesissi@szu.edu.cn). H.-T. Wai and A. Scaglione are with the Ira A. Fulton School of Electrical Computer and Energy Engineering, Arizona State University, USA (e-mail: htwai.Scaglione@asu.edu; Anna.Scaglione@asu.edu). L. Li is with the Massachusetts Institute of Technology Lincoln Laboratory, USA (e-mail: lin.li@ll.mit.edu).
Publisher Copyright:
© 1963-2012 IEEE.
PY - 2018/8
Y1 - 2018/8
N2 - Principal component analysis (PCA) is a fundamental primitive of many data analysis, array processing, and machine learning methods. In applications where extremely large arrays of data are involved, particularly in distributed data acquisition systems, distributed PCA algorithms can harness local communications and network connectivity to overcome the need of communicating and accessing the entire array locally. A key feature of distributed PCA algorithm is that they defy the conventional notion that the first step toward computing the principal vectors is to form a sample covariance. This paper is a survey of the methodologies to perform distributed PCA on different data sets, their performance, and of their applications in the context of distributed data acquisition systems.
AB - Principal component analysis (PCA) is a fundamental primitive of many data analysis, array processing, and machine learning methods. In applications where extremely large arrays of data are involved, particularly in distributed data acquisition systems, distributed PCA algorithms can harness local communications and network connectivity to overcome the need of communicating and accessing the entire array locally. A key feature of distributed PCA algorithm is that they defy the conventional notion that the first step toward computing the principal vectors is to form a sample covariance. This paper is a survey of the methodologies to perform distributed PCA on different data sets, their performance, and of their applications in the context of distributed data acquisition systems.
KW - Clustering algorithms
KW - data mining
KW - distributed algorithms
KW - principal component analysis
KW - radar signal processing
UR - http://www.scopus.com/inward/record.url?scp=85051204310&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85051204310&partnerID=8YFLogxK
U2 - 10.1109/JPROC.2018.2846568
DO - 10.1109/JPROC.2018.2846568
M3 - Article
AN - SCOPUS:85051204310
SN - 0018-9219
VL - 106
SP - 1321
EP - 1340
JO - Proceedings of the IEEE
JF - Proceedings of the IEEE
IS - 8
M1 - 8425655
ER -