.png){kind=icon}
.png){kind=icon}
Return
A Kernel Normalized Data-Reusing Generalized Maximum Correntropy Algorithm
.png){kind=icon}
PRE
AI
DOI：10.1109/TCSII.2024.3361812.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The kernel adaptive filtering (KAF) algorithms have drawn increasing attention due to their advantages such as universal nonlinear approximation, linearity, and convexity in reproducing kernel Hilbert space (RKHS). This fact motivates us to develop a new KAF algorithm called kernel normalized data-reusing generalized maximum correntropy (KNDR-GMC). The cost function for KNDR-GMC is constructed by combining the generalized correntropy criterion, data-reusing method, and Gaussian kernel to deal with nonlinear system identification under impulsive noisy environments. Compared with existing algorithms, in a chaotic time-series prediction and a real-world application, the simulation results verify that KNDR-GMC achieves better filtering accuracy and a faster convergence rate.
Keywords:
Kernel
Cost function
Prediction algorithms
Noise measurement
Convergence
Approximation algorithms
Density functional theory
Kernel adaptive filtering
generalized correntropy
data-reusing
impulsive noise
Journal
I
IF:
4.9
Papers:
8.8K
Citations:
2.5W