Finding minima

I have a data file with few valleys (minima). is there a way to find all the x coordinates corresponding to each minimum y value?

Answer 1

[Matt Tearle]

If you want fancy, take Walter's advice and search FEx. If you want quick and dirty:

x = 1:50;
y = sin(5*x);
idx = [false, y(3:end)>y(2:end-1) & y(2:end-1)<y(1:end-2), false];
xmin = x(idx)
plot(x,y,'o-',xmin,y(idx),'rx')

[Walter Roberson]

The above code unfortunately picks up on every local valley; the contributions in the FEX worry about noise and defining how much of a change there can be before saying that one has entered a new extreme or if two adjacent extreme should be considered part of the same series or not...

[Matt Tearle]

Right, that's why I said "quick and dirty"

[Paulo Silva]

Matlabs findpeaks will do the job just fine, the recent matlab version have a better implementation of findpeaks.

[Matt Tearle]

True, but note: FINDPEAKS is a Signal Processing Toolbox function.

[Paulo Silva]

Matt your simple detection is very interesting :) I never thought about it that way, it's not perfect but might be very useful, thanks for sharing it.

[Matt Tearle]

Hah. Actually I just hunt the intartubes looking for places to apply logical indexing! ;)

[Paulo Silva]

Here's my 1 vote :)

Answer 2

[Walter Roberson]

There is are a fair number of postings, and several Matlab File Exchange contributions, for finding peaks, which you could either modify or apply to the negative of your data.

Answer 3

[Andrew Newell]

There is also the Optimization Toolbox.

Answer 4

[Graig]

Thanks All. For me actually the 'quick and dirty' way works well, as the data are very smooth (not noisy). Seemed to me, this is what the black box 'findpeaks' contains in a better way.

[Matt Tearle]

To help others who might have the same question, it would be great if you could vote for and/or accept answers that you think should be given prominence. Thanks.