Comparison of Least-Squares and Stochastic Gradient Lattice Predictor Algorithms Using Two Performance Criteria

Michael L. Honig, David G. Messerschmitt

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

The least-squares (LS) and stochastic gradient (SG) lattice prediction algorithms are compared using two different performance criteria. These are a) output mean squared error and b) the accuracy of the autoregressive spectral estimate obtained from the mean values of the lattice coefficients, assuming a stationary input. It is found that the second performance criterion is more sensitive than the first. This “spectral” performance criterion is a measure of the accuracy of the estimated autoregressive model coefficients. Bias in the LS and SG coefficient estimates can cause significant deviation of the asymptotic spectral estimates from the actual input spectrum: The similarly between the LS and SG lattice algorithms enables comparative simulations.

Original languageEnglish (US)
Pages (from-to)441-445
Number of pages5
JournalIEEE Transactions on Acoustics, Speech, and Signal Processing
Volume32
Issue number2
DOIs
StatePublished - 1984

ASJC Scopus subject areas

  • Signal Processing

Cite this