Performing Discriminant Analysis
Using the Discriminant Platform
Discriminant Analysis seeks to find a way to predict a classification (X) variable (nominal or ordinal) based on known continuous responses (Y). It can be regarded as inverse prediction from a multivariate analysis of variance (MANOVA). In fact, the Manova personality of Fit Model also provides some discriminant features.
• A stepwise selection dialog to help choose variables that discriminate well.
• Choice among Linear, Quadratic, or Regularized-Parameter analyses.
• The discriminant scores, showing which groups each point is near.
• A misclassification summary.
• Plots of the points and means in a canonical space, which is the space that shows ...