0

I am running an LDA model on the iris dataset for two class (any two).

I want to now calculate the Fisher discriminant value for Fisher projection.

Does this mean I need the svd value?

ttnphns
  • 51,648
  • 40
  • 253
  • 462
Hardik Gupta
  • 161
  • 1
  • 1
  • 7
  • Are you speaking of discriminant functions values or of Fisher's classification functions values? (See footnote here: https://stats.stackexchange.com/a/190821/3277 about terminologic confusions) – ttnphns Aug 19 '17 at 16:38
  • Fisher's classification functions because the task is for classification – Hardik Gupta Aug 19 '17 at 16:42
  • OK then, there in the answer is the link to another answer showing formulas how those can be computed (at least, SPSS does so, I believe other programs are doing same way). – ttnphns Aug 19 '17 at 16:44
  • Ah, here is the direct link. Look under "Addition". https://stats.stackexchange.com/a/31384/3277 – ttnphns Aug 19 '17 at 16:45
  • Actually the confusion is all I am asked to do is calculate the Fisher discriminant value for Fisher projection. There is no further explaination – Hardik Gupta Aug 19 '17 at 17:05
  • But this time you sound like you want to compute discriminant function values. For 2 class aka Fisher's LDA. There is a number of answers on this site showing it. Did you try a search? – ttnphns Aug 19 '17 at 17:49

0 Answers0