Skip to main content
1-Visitor
December 4, 2010
Solved

peaks for various graphs?

  • December 4, 2010
  • 3 replies
  • 7902 views

Hi,

im having trouble with finding the peaks of a various data sets.

im attaching 3 data sets.. is there any way that the peaks can be found for them?

im thinking of smoothing the data first using supsmooth and then it will be easier. so im attaching the matcad 11 file and im also adding some explanation within it.

update:

searched google and came accross this. its one of the sample files

http://support.mathsoft.com/support/mcad8/Examples/Locate_peaks.mcd

but its not there anymore.. does anyone know where can I get it? or is it even what im looking for

Message was edited by: UKMan

Best answer by RichardJ

Here's another approach. I've posted these functions before I think, but I don't recall which thread (and can't be bothered to search for it!). These are newer versions though, so if anyone has the old ones they should replace them.

As a note, if you choose the window as 10 and the threshold as 5% it will find the two peaks in the unsmoothed data, in all three cases. For data set 1 it finds only the higher peak in the doublet. If you actually wanted to find both peaks in the doublet you could change the window size to 4.

3 replies

23-Emerald V
December 5, 2010

very primitive method that relies upon having reasonably smooth data. haven't got time to find a 'proper' method, so this is just a top of the head thing.

Stuart

23-Emerald V
December 5, 2010

Of course, if I had any brains then apart from using a proper technique, I would have embedded the smoothing function within the peak detector.

Stuart

1-Visitor
December 5, 2010

Lovely work, great explanatory worksheet.

Mike

RichardJ19-TanzaniteAnswer
19-Tanzanite
December 5, 2010

Here's another approach. I've posted these functions before I think, but I don't recall which thread (and can't be bothered to search for it!). These are newer versions though, so if anyone has the old ones they should replace them.

As a note, if you choose the window as 10 and the threshold as 5% it will find the two peaks in the unsmoothed data, in all three cases. For data set 1 it finds only the higher peak in the doublet. If you actually wanted to find both peaks in the doublet you could change the window size to 4.

1-Visitor
December 5, 2010

Very nice Richard.

There would be a neat method if the number of peaks is known. The data could be sorted and the corresponding peak values could then be extracted.

Simple but effective

Mike

24-Ruby IV
December 6, 2010

A have calculated by Mathcad help a spesific vavue of water and steam by 2 formulations - IAPWS-95 and IAPWS-IF97

The different between this calculations you can see as a set os peaks: see surface and counter plot