[R] Sensitivity and specificity in multi-class problems

Schäfer Christian c.schaefer at dkfz-heidelberg.de
Sat May 19 21:38:07 CEST 2007


I'm relative new to R.
I'm looking for a convenient way to calculate sensitivity and
specificity in machine-learning models (SVM, ...). In a two-class case
it's quite simple by calculating the TP, TN, FP, FN rates once but in the
multi-class case I have to do it as often as there exist different
classes (for each class).
Of course I could do it manually in a for-loop or so. But is there a
more convenient way? Perhaps on the basis of a confusion table?

Thanks in advance.

Christian Schaefer

More information about the R-help mailing list