Trying to find out the performance of DLMS based detection. Before that I need check out the stability of estimation performance. But I don't know the way how to check the stability of estimation performannce.
Hi,
K-fold estimation is useful. the data should be divided to K subset. Accuracy and error of estimation regarding each subset are evaluated. Then the standard deviation of them is an indicator of stability.
Regards, Mahdieh
The convergence of the covariance to a finite steady state — that is, the error
becoming a stationary process in the MS sense — is equivalent to filter
stability in the bounded input bounded output sense.
Stability of the filter does not require the dynamic system to be stable —only
the observability condition is required. The observability
alone does not guarantee uniqueness — the steady-state solution might depend
on the initial covariance — but the existence (finiteness) of the solution is the
key.
You can use the normalized estimation error squared (NEES).
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis