A MEASURE OF ASSOCIATION FOR ASSESSING PREDICTION ACCURACY THAT IS A GENERALIZATION OF NONPARAMETRIC ROC AREA

There is a need for a measure of prediction accuracy that generalizes non-parametric receiver operating characteristic (ROC) area to polytom ous ordinal patient state. We describe such a measure, prediction prob ability P-K, derived from Kim's measure of association. We show that t he value of P-K equals the value of non-parametric ROC area for dichot omous patient state and is a meaningful generalization of nonparametri c ROC area for polytomous state.