TY - JOUR
T1 - Robust learning with Kernel mean p-power error loss
AU - Chen, Badong
AU - Xing, Lei
AU - Wang, Xin
AU - Qin, Jing
AU - Zheng, Nanning
N1 - Funding Information:
Manuscript received December 20, 2016; revised May 20, 2017; accepted July 9, 2017. Date of publication July 25, 2017; date of current version June 14, 2018. This work was supported in part by the 973 Program under Grant 2015CB351703 and in part by the National NSF of China under Grant 91648208 and Grant 61372152. This paper was recommended by Associate Editor G.-B. Huang. (Corresponding author: Badong Chen.) B. Chen, L. Xing, X. Wang, and N. Zheng are with the Institute of Artificial Intelligence and Robotics, Xi’an Jiaotong University, Xi’an 710049, China (e-mail: [email protected]; [email protected]; [email protected]; [email protected]).
Publisher Copyright:
© 2013 IEEE.
PY - 2018/7
Y1 - 2018/7
N2 - Correntropy is a second order statistical measure in kernel space, which has been successfully applied in robust learning and signal processing. In this paper, we define a nonsecond order statistical measure in kernel space, called the kernel mean-p power error (KMPE), including the correntropic loss (C-Loss) as a special case. Some basic properties of KMPE are presented. In particular, we apply the KMPE to extreme learning machine (ELM) and principal component analysis (PCA), and develop two robust learning algorithms, namely ELM-KMPE and PCA-KMPE. Experimental results on synthetic and benchmark data show that the developed algorithms can achieve better performance when compared with some existing methods.
AB - Correntropy is a second order statistical measure in kernel space, which has been successfully applied in robust learning and signal processing. In this paper, we define a nonsecond order statistical measure in kernel space, called the kernel mean-p power error (KMPE), including the correntropic loss (C-Loss) as a special case. Some basic properties of KMPE are presented. In particular, we apply the KMPE to extreme learning machine (ELM) and principal component analysis (PCA), and develop two robust learning algorithms, namely ELM-KMPE and PCA-KMPE. Experimental results on synthetic and benchmark data show that the developed algorithms can achieve better performance when compared with some existing methods.
KW - Extreme learning machine (ELM)
KW - kernel mean p-power error (KMPE)
KW - principal component analysis (PCA)
KW - robust learning
UR - http://www.scopus.com/inward/record.url?scp=85028918804&partnerID=8YFLogxK
U2 - 10.1109/TCYB.2017.2727278
DO - 10.1109/TCYB.2017.2727278
M3 - Journal article
C2 - 28749366
AN - SCOPUS:85028918804
SN - 2168-2267
VL - 48
SP - 2101
EP - 2113
JO - IEEE Transactions on Cybernetics
JF - IEEE Transactions on Cybernetics
IS - 7
ER -