Nonparametric Monotone Classification with MOCA

We describe a monotone classification algorithm called MOCA that attempts to minimize the mean absolute prediction error for classification problems with ordered class labels. We first find a monotone classifier with minimum L-1 loss on the training sample, and then use a simple interpolation scheme to predict the class labels for attribute vectors not present in the training data. We compare MOCA to the Ordinal Stochastic Dominance Learner (OSDL), on artificial as well as real data sets. We show that MOCA often outperforms OS D L with respect to mean absolute prediction error\