3
$\begingroup$

I'm trying to get an estimate of the error in my sample standard deviation (not a statistician at all, hope I'm using correct terminology). The timeseries is normally distributed and follows an AR1 process to good enough approximation.

I understand how to do it without autocorrelation (Standard deviation of standard deviation) and how to compute a correction for AR1 autocorrelation in an estimate of standard error for an estimate of the mean. (https://en.wikipedia.org/wiki/Standard_error#Correction_for_correlation_in_the_sample), but I'm uncertain whether I can just add this correction to the estimate of standard deviation of standard deviation.

(if you know of a reference, that would also be much appreciated)

Improve this question
$\endgroup$
1
  • $\begingroup$ Explicit simulation of an AR1 process leads me to believe you can indeed simply compute the standard deviation of standard deviation by this correction factor. Don't trust it completely and still looking for a reference... $\endgroup$
    – Femkemilene
    Commented Aug 22, 2018 at 14:04

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.