Help me determine why the Isolocalmax doesn't work after applying MinProminence.

 % Identify indirect peak values (aka vallies)
                dataDirect = dataOut;
                x = 1: (length(dataDirect));
                [ TFm, Pm] = islocalmax(dataDirect);
                min_localmax = (mode(dataDirect(TFm))+ mean(dataDirect(TFm)))/2;
%                 max2 = max(dataDirect);
%                 max_minlimit2 = max2*.9990;
                [ TFd, Pd] = islocalmax(dataDirect, 'MinProminence', min_localmax);% max_minlimit2); % aveDiff,
                                
                Peak_location1 = (x(TFd))'; 
                Peak_values1 = dataDirect(TFd);


                nexttile
                hold on
                yline(min_localmax,'r-')
                plot(x, dataDirect,x(TFd),dataDirect(TFd),'r*')
                title('butter ')

For some reason, the islocalmax function is not identifying any peaks when I use 'MinProminence.' If anyone could suggest alternative functions that would be great. The 'islocalmax' works well when I am trying to determine by 'MinProminence' value but after that it no longer works and I can't figure out why since I use this function all of the time.

If anyone has the time to help I would really appreciate it!

Answer 1

[Star Strider]

You need to check the variable values.

I cannot run this here (the Run feature is currently not working, at least for me, although MATLAB Online works), however running this:

dataOut = randn(500,1);

dataDirect = dataOut;

x = 1: (length(dataDirect));

[ TFm, Pm] = islocalmax(dataDirect);

min_localmax = (mode(dataDirect(TFm))+ mean(dataDirect(TFm)))/2

datamax = max(dataOut) % Added

minPM = min(Pm(Pm ~= 0 ))

produces this result:

min_localmax =

-0.2613

datamax =

3.5784

minPM =

0.0082

So I suspect using an extremely low value for 'MinProminence' (much less than the lowest prominence value returned) is likely the problem.

EDIT — (31 Mar 2024 at 17:00)

Corrected typographical errors.

.