How pca() function works in matlab

1 view (last 30 days)
I'm trying to apply pca() function on 3d data, But I'm getting empty COEFF matrix
x1 = randn(400,1);
y1 = randn(400,1);
z1 = randn(400,1);
plot3(x1,y1,z1,'r.');
[COEFF] = pca([x1;y1;z1]')

Accepted Answer

Jason Millen
Jason Millen on 16 Aug 2016
Edited: Jason Millen on 16 Aug 2016
The reason you are getting an empty 1200-by-0 matrix is because you are passing a 1-by-1200 matrix to pca.
I believe what you want to do is pass it a 400-by-3 matrix, which can be accomplished with [x1';y1';z1']'. Thus, if you call pca([x1';y1';z1']'), you will get a 3-by-3 coefficient matrix.
If you would like to read up more on the documentation for pca(), you can find it here.
  3 Comments
Jason Millen
Jason Millen on 16 Aug 2016
I believe there is a transpose operator given before the matrix is passed into pca, making it 1-by-1200. Please correct me if I am wrong.
Walter Roberson
Walter Roberson on 16 Aug 2016
Ah, you are correct, I missed that.

Sign in to comment.

More Answers (0)

Categories

Find more on Dimensionality Reduction and Feature Extraction in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!