Category "non-linear"

Choosing number of dimensions in LLE

Is there a way to choose the number of reduced dimensions d where d<D and D is the number of original dimensions in the original dataset? I'm trying to use t