Help using Base SAS procedures

kNN using PROC DISCRIM

Reply
Occasional Contributor
Posts: 5

kNN using PROC DISCRIM

Hi All -

In view of a tie in the estimated posterior probabilities between two or more populations, a particular observation is classified as OTHER when performing kNN with PROC DISCRIM.

Isn’t there a difference in truly misclassifying an observation into an incorrect population versus having a tie thus classifying the observation as OTHER?

Based on the current output of the total error count for the training data set, those two situations are treated the same.

Any thoughts on this matter would be much appreciated.

Thanks,

Song
Ask a Question
Discussion stats
  • 0 replies
  • 138 views
  • 0 likes
  • 1 in conversation