We collect the ECG signal by ourselves and have already detected the R peaks by our method, but we don't have any reference signal to verify the R peaks we found are accurate.
It seems like the statistical method such as sensitivity, positive predictivity, the error rate are the improper way to use.
Can someone please suggest some statistical method to evaluate the performance of peak detection without a reference signal.
Like, calculate the average RR interval?
Thank you all for your opinions. I highly appreciate
Fernando Soares Schlindwein
Get an expert to generate the 'gold standard'. Better still, 3 experts and implemment a vote of off one out and averaging the agreeing 2. Yes, time consumining and expensive. Certain things simply cannot be done 'on the cheap'.
0 votes
0 thanks
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude
Similar questions and discussions