Title :
Analysis of linear prediction by matrix approximation
Author :
Kot, Alex C. ; Tufts, Donald W. ; Vaccaro, Richard J.
Author_Institution :
Nanyang Technol. Inst., Singapore
fDate :
11/1/1993 12:00:00 AM
Abstract :
The authors present a new method of analysis of the effects of data perturbations on the estimates of signal parameters which are obtained by the method of linear prediction. The variances of signal-parameter obtained from linear prediction methods are computed. Although the results appear to have wide applicability, for simplicity the authors restrict their examples to signals which are linear combinations of complex-valued exponential functions
Keywords :
filtering and prediction theory; matrix algebra; parameter estimation; backward error analysis; complex-valued exponential function; data perturbations; linear prediction; matrix approximation; signal parameter estimation; Error analysis; Frequency estimation; Linear approximation; Matrix decomposition; Noise level; Parameter estimation; Prediction methods; Signal analysis; Singular value decomposition; System identification;
Journal_Title :
Signal Processing, IEEE Transactions on
