I am trying to use the pm-prophet feature auto_changepoints to detect changepoints in a dataset. It gives seemingly good results, but when running the script on the same dataset twice, the inferred changepoints are located different places. How can i make the auto_changepoints more accurate? does this have to do with my sample size?
You can probably get better support by asking on their Github issue, but my guess is that you need stronger prior on the change point location.