Message Boards Message Boards

FindPeaks having issues with TimeSeries

Posted 4 months ago

Hi,

FindPeaks does not like something about this TimeSeries object. I have run it on other TimeSeries and it completes without errors. Any ideas what is going on here. Example is in the attached notebook.

Thanks,

POSTED BY: Edmund Robinson
3 Replies
Posted 1 day ago

While it's annoying to get error messages that aren't easily interpreted, if your data was from biological data that I typically deal with, I'd say you have no peaks as the minor fluctuations that appear look more like random noise.

Are the measures of precision for each data point or some overall summary of the measurement error small enough to justify searching for peaks?

POSTED BY: Jim Baldwin

My experience is that the default value of the σ argument in FindPeaks can be too large. If so, then that's why forcing σ to 0 helps.

Unfortunately, 0 may not always be an appropriate value, so I adjust σ as needed, with this limitation in mind. Sometimes no satisfactory value will work.

I consider this to be a serious bug in FindPeaks. This function either shouldn't fail with larger values of σ, or it should fail more gracefully. Instead, the circumstances of failure can seem somewhat arbitrary, and the resulting error message is obtuse and unhelpful.

POSTED BY: Robert Rasmussen

Example from help pages of FindPeaks -> Scope -> Data -> Second example:

POSTED BY: Mariusz Iwaniuk
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract