TY - JOUR
T1 - Stochastic information gradient algorithm with generalized gaussian distribution model
AU - Chen, Badong
AU - Principe, Jose C.
AU - Hu, Jinchun
AU - Zhu, Yu
PY - 2012/2
Y1 - 2012/2
N2 - This paper presents a parameterized version of the stochastic information gradient (SIG) algorithm, in which the error distribution is modeled by generalized Gaussian density (GGD), with location, shape, and dispersion parameters. Compared with the kernel-based SIG (SIG-Kernel) algorithm, the GGD-based SIG (SIG-GGD) algorithm does not involve kernel width selection. If the error is zero-mean, the SIG-GGD algorithm will become the least mean p-power (LMP) algorithm with adaptive order and variable step-size. Due to its well matched density estimation and automatic switching capability, the proposed algorithm is favorably in line with existing algorithms.
AB - This paper presents a parameterized version of the stochastic information gradient (SIG) algorithm, in which the error distribution is modeled by generalized Gaussian density (GGD), with location, shape, and dispersion parameters. Compared with the kernel-based SIG (SIG-Kernel) algorithm, the GGD-based SIG (SIG-GGD) algorithm does not involve kernel width selection. If the error is zero-mean, the SIG-GGD algorithm will become the least mean p-power (LMP) algorithm with adaptive order and variable step-size. Due to its well matched density estimation and automatic switching capability, the proposed algorithm is favorably in line with existing algorithms.
KW - Minimum error entropy criterion (MEE)
KW - generalized Gaussian density (GGD)
KW - least mean p-power (LMP)
KW - stochastic information gradient (SIG) algorithm
UR - https://www.scopus.com/pages/publications/84863272885
U2 - 10.1142/S0218126612500065
DO - 10.1142/S0218126612500065
M3 - 文章
AN - SCOPUS:84863272885
SN - 0218-1266
VL - 21
JO - Journal of Circuits, Systems and Computers
JF - Journal of Circuits, Systems and Computers
IS - 1
M1 - 1250006
ER -