Area between curve (numerical data with positive and negative values) and y = 0

Hi all,

I am trying to calculate the area between a curve (defined by numerical data rather than a function) and the x-axis. I don't think it is a straightforward trapz application as the curve has noise that at times oscillates above and below y = 0 (i.e. the numerical data is positive and negative) and I don't want the 'negative' areas to be subtracted from the positive ones. Hopefully that makes sense...

An example of the data is attached.

Thank you so much in advance!

Edit: To clarify, I don't want to areas under y = 0 to contribute to the result at all so I can't abs(y). Thanks again!

Answer 1

[Star Strider]

I do not have a clear idea of what you want as the result.

It is relatively straightforward to integrate the areas greater than 0 and the areas less than 0 separately:

x = 0:100;
y = randn(1, 101);
gt0 = y>0;

posArea = trapz(x(gt0), y(gt0));                        % Area > 0
negArea = trapz(x(~gt0), y(~gt0));                      % Area < 0

If you want greater resolution in the areas, use interp1 first (before doing the logical comparisons and trapz calls) to interpolate to more detailed independent and dependent variable values.

[Fiona]

I can't believe I didn't figure that out myself. I was making it way too hard for myself! Thank you so much!

[Star Strider]

As always, my pleasure!