Thursday, October 27, 2011

10/25/2011

When LSI is done on classification problems, we pick dimensions that maximally differentiate across classes,
while having minimal variance with in any given class.This method is called as LDA - Linear Discriminant Analysis.

-Sandeep Gautham