Why can't I use the PCA function on my 513*40 matirx with positive and negative values?

I have a 513*40 matrix of EEG data so that means microvolt values ranging from -30 to 30. I want to use the function PCA in order to reduce the dimensionality so I can use a pattern recognition software but when I say coeff=pca(x) it gives me the error message that subscript must be positive real integers or logicals. I have tried taking the RMS of the matrix to make all positive but I still get the same error. Can anyone explain why?

anxious for reply. Thanks Kim

Answer 1

[Image Analyst]

You have an array in your program called pca. This causes it to take precedence over the toolbox function. So it's assuming x are indexes of your array called pca instead of inputs to the function pca. Rename your pca array to something else, like pcaCoefficients or something.

[Kim]

Thank you I' will make sure I don't have anything called that!