How to calculate the variance of linear prediction parameters?

Question

I'm using linear prediction with singular value decomposition (LPSVD) to analyze signals that are damped sinusoids. If my understanding of the theory of linear prediction is correct (and it may not be) then LPSVD is actually doing a linear least squares fit. If this is correct then I should be able to calculate the estimated variance in the model parameters in a way similar to a simple linear regression of one variable.

Can anyone suggest a reference that might have useful information? Thanks.

Answer 1

It seems like the best way to calculate the variance on the parameters returned by the LPSVD algorithm is to use the Cramér-Rao bound as the estimate. If anyone has a more elegant solution I'd like to hear it.