Principle Component Analysis - Problem in Example

Hi,

I've been learning about principle component analysis recently in order to reduce data for a fault classification problem. I'v referred to the worked example here to try and aid my understanding of it:

http://www.mathworks.co.uk/help/stats/feature-transformation.html#f75476

However, when I attempt to use the pca function with the data matrix and the weights as input, an error is thrown saying that it does not accept the weights, w, as an input (error: too many input arguments). I may be missing something obvious here, but has anyone got any ideas to what the problem is?

Thanks

Rory

Answer 1

[Shashank Prasanna]

Hi Rory, What version of MATLAB are you using? PCA (with supported for wighted and missing values) was introduced only in R2012b:

http://www.mathworks.com/help/stats/release-notes.html

You may try the following to confirm that you have the right PCA:

>> which -all pca

[Rory]

Hi Shashank,

Thanks for replying. I realised just before your answer that this is probably a new feature for the PCA function on R2012b. I'm running R2011b. Is it possible to update the toolbox?

Rory

[Walter Roberson]

The only way to update the toolbox is to also update MATLAB.

[Shashank Prasanna]

Rory, you will have to update your MATLAB. You won't be able to just update the toolbox. You may want to contact installation support if you have SMS.