0

I think the lda function in R's MASS library uses the Mahalanobis distance function. Is it possible to force the lda function to use Euclidean distance and Manhatten distance instead? It would make it much easier for students to visualize / verify the distances to the various centroids and therefore the resulting classification. Thank you.

Bill Qualls
  • 397
  • 1
  • 7
  • 20

0 Answers0