Trevor Hastie and Robert Tibshirani,
Classification by Pairwise Coupling.
Stanford University and University of Toronto, November 1996.
Trevor Hastie and Robert Tibshirani,
Classification by Pairwise Coupling.
Stanford University and University of Toronto, November 1996.
We discuss a strategy for polychotomous classification that involves estimating class probabilities for each pair of classes, and then coupling the estimates together. The coupling model is similar to the Bradley-Terry method for paired comparisons. We study the nature of the class probability estimates that arise, and examine the performance of the procedure in real and simulated datasets. Classifiers used include linear discriminants, nearest neighbours, and the support vector machine.