PCA

So I decided to put LDA on the back burner because it was taking too much time (see the previous posts).  Last night I started to implement PCA and 5 minutes ago I finished it.  It works well.  It can reduce the swiss roll data set in next to no time (which is what you’d expect).  By having this implementation of PCA it means I can carry on with other areas of the dissertation and then given time come back and implement LDA and Isomap.