How to use Principal Component Analysis to reduce feature vector size?

I am working on character recognition. Feature vector size i got is 200x1 How can i use pca analysis to reduce the feature vector ?

Answer 1

[Image Analyst]

Run it through pca(). Look at the first few components and their loadings/weightings. If some feature is not being used by the components you want to use, then you can try getting rid of it. If your classifications are the same, then those features were not really needed.