Title :
On the normalization of the step size in nonsymmetric stochastic gradient algorithms
Author_Institution :
Eurecom Inst., Valbonne, France
Abstract :
The normalization of the step size in the least-mean-square (LMS) algorithm allows easy control of the range of stable operation for the normalized step size in the normalized LMS (NLMS) algorithm, and also easy determination of the step size for maximum convergence speed. A generalized step size normalization for stochastic gradient algorithms in which the gradient vector differs from the data vector is proposed. Three applications are considered in detail: the stochastic Newton scheme, the sign-data LMS algorithm, and an instrumental variable method proposed to speed up the convergence of the LMS algorithm
Keywords :
adaptive filters; convergence; digital filters; filtering and prediction theory; least squares approximations; stochastic processes; FIR filters; least mean square algorithm; maximum convergence speed; nonsymmetric stochastic gradient algorithms; stable operation; step size normalisation; stochastic Newton scheme; Algorithm design and analysis; Convergence; Eigenvalues and eigenfunctions; Equations; Finite impulse response filter; Instruments; Least squares approximation; Size control; Statistics; Stochastic processes;
Conference_Titel :
Signals, Systems and Computers, 1992. 1992 Conference Record of The Twenty-Sixth Asilomar Conference on
Conference_Location :
Pacific Grove, CA
Print_ISBN :
0-8186-3160-0
DOI :
10.1109/ACSSC.1992.269147
