Context:
When performing a discriminant analysis in JMP, the score Summaries give a number of "Misclassified observations". The user manual pdf "multivariate methods" define the following:
Number Misclassified: Provides the number of observations in the specified set that are incorrectly classified.
Percent Misclassified: Provides the percent of observations in the specified set that are incorrectly classified.
Prob(Pred): Estimated probability of the observation’s predicted classification.
Question
It is unclear to me on what probability is an observation is well classified or not. In other words, which this the true Equation behind the "misclassified"?
Possible answers?
1) if Prob(Pred) defined the well classified or not, how is it calculated? For instance, (when looking at the "saved formula" section of the multivariate analysis manual), is it calculated from "qt prior probability of membership for group t " or "the p(t|y) posterior probability that y belongs to group t"
Thanks for your answers
PS: the question was edited for more clarity
For current version (13), select Help > Books > Multivariate Methods. See Chapter 5: Discriminant Analysis. See Saved Formulas in the last section, Technical Details.
Earlier versions should have the same help in a similar path.
Discriminant analysis is a form of supervised learning. You provide a known response with the training data set. The discriminant analysis results in a predictive model. The predicted response is compared to the given response. If they don't match, then that observation is considered misclassified.
