Alternating between if conditions in a loop.

I am trying to write code to isolate repeating patterns in a dataset. The example I will give is an oscillatory dataset to isolate peaks and troughs, and save the values within vectors.

My raw data looks like this:

I can fit a poor model to this dataset to get:

from the reproducable model:

model = @(p, x) p(1)*2*pi*sin(2*pi*p(2)*x +p(3));
params = [3.6e+03 46.5 1000];
x = linspace(1.5, 2.5, 100000);
y = model(params, x);

As can be seen from the first figure, the peaks and troughs occur at different amplitudes, and the green model does not refect this.

So what I want to do is to loop through the y-data presented in the first figure to find when the value at each peak and trough. To do this I was thinking about using a for - loop to run through the entire dataset, and use conditions to locate the values of the peaks:

itr = length(y);
for i = 2:itr
    a = y(i) - y(i-1);
    if a < 0
        %Save a%
        Max(1) == y(i)
    end
end

This will save the maximum value of y from the first oscillation period. From this point onwards, a = y(i) - y(i-1); will be negative until y(i) reaches the minima. So after saving Max(1), I would like to switch the condition to if a > 0, to find when a becomes positive and then save y(i) as a new object, Min(1), and to then revert back to the original condition of if a < 0. I would like to repeat this process until i = itr, appending maximum and minimum y values to the objects Max and Min, respectively.

I'd like to achieve this without using inbuilt functions if possible

[Star Strider]

I am not dertain what you want to do. I would use either findpeaks (with normal and negated arguments to get the peaks and troughs respectively) or islocalmax and islocalmin to do the same. These funcitons already exist and are optimised, so much more efficient than repeated explicit loops.

[James]

Hi Star Strider. I would rather avoid inbuilt functions, as the methods that I described could be useful for other processes. I'm trying to ascertain how to to gain greater control over loops and condition statements.

Answer 1

[James]

Okay, so there's a really simple solution that requires a condition to be pre-set:

itr = length(y_1);
Max = [];  % Initialize empty array to store maximum values
Min = [];  % Initialize empty array to store minimum values
maxCondition = true;  % Flag to indicate the current condition
MInd1 = []; % Maximum indeces
MInd2 = []; % Minimum Indeces
for i = 2:itr
    a = y_1(i) - y_1(i-1);
    
    if maxCondition && a < 0
        Max(end+1) = y_1(i-1);  % Append maximum value to Max array
        maxCondition = false;   % Switch condition to look for minimum
        MInd1(end+1) = i-1;     % Append indeces array for the maximum values
    elseif ~maxCondition && a > 0
        Min(end+1) = y_1(i-1);  % Append minimum value to Min array
        maxCondition = true;    % Switch condition to look for maximum
        MInd2(end+1) = i-1;     % Append indeces array for the minumum values
    end
end

Here we set two conditions that correspond to whether searching for a maximum or a minimum value I know from the data I am searching for a maximum value so I pre-set it to 'true'.

The conditions are altered if the if statement is triggered, so when a maximum value is found, the maxCondition is set to false.

Here are the charts to show the maxima and minima have been found: