how to run principal component analysis in a 3D matrix

Hi,

I am trying to run principal component analysis, pca() function to a 3D matrix. It does not work and I think it only works with 2D matrixes. Is there any way to circunvent this limitation?

Thank you,

Best regards,

Answer 1

[AndresVar]

You can reshape the matrix to 2D and then when you get results convert it back to the orginal dimensions if needed

Reshape array - MATLAB reshape (mathworks.com)

There tricky part is to choose how to reshape. But say 2 dimensions are data, and the third is time then

data3d = ones([2,2,3]);
data2d = reshape(data3d,[],size(data3d,3));
size(data2d)

ans = 1×2

4 3

so the columns become the new time dimension.

Answer 2

[Image Analyst]

See my attached demo where I run it on a 3-D (true color) image. Adapt as needed.