Learning characteristics of stochastic-gradient-descent algorithms: A general study, analysis, and critique

Signal Processing

Volumn 6, Issue 2, 1984, Pages 113-133

  Gardner, W A ^a  

^a UNIVERSITY OF CALIFORNIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

SIGNAL PROCESSING - ANALYSIS;

SIGNAL FILTERING AND PREDICTION;

EID: 0021410050     PISSN: 01651684     EISSN: None     Source Type: Journal    
DOI: 10.1016/0165-1684(84)90013-6     Document Type: Article

References (23)

1. Recursive least squares ladder estimation algorithms
2. Adaptive equalization of highly dispersive channels for data transmission
3. Theory on the speed of convergence in adaptive equalizers for digital communication
4. On the design of gradient algorithms for digitally implemented adaptive filters
5. Adaptive noise cancelling: Principles and applications
6. Stationary and nonstationary learning characteristics of the LMS adaptive filter
7. The effects of large interference on the tracking capability of digitally implemented echo cancellers
8. Adaptive enhancement of multiple sinusoids in uncorrelated noise
9. On the required tap-weight precision for digitally implemented, adaptive, mean-squared equalizers
10. On the independence theory of equalizer convergence
11. Stochastic approximation with correlated data
12. Tracking properties of adaptive signal processing algorithms
13. Analysis of error-gradient adaptive linear estimators for a class of stationary dependent processes
14. Second-order output statistics of the adaptive line enhancer
15. Tracking characteristics of the LMS adaptive line enhancer-response to a linear chirp signal in noise
16. Linear estimation in an unknown quasi-stationary environment
17. Adaptive linear discrete-time estimation
18. Performance advantage of complex LMS for controlling narrow-band adaptive arrays
19. On the asymptotic eigenvalue distribution of Toeplitz matrices
20. Toeplitz and circulant matrices: II