0

I am trying to reduce the size of my features vectors using PCA and LDA. Following the approach presented here, I cannot understand step 3 and step 4 described in that approach. Why is the multiplication from step 3 needed and why the projection is not done on LDA eigenvectors?

That isn't really an AI question; have you tried a more stats-related site? – Oliver Mason – 2020-03-30T08:33:05.613