The main objective of this paper is to propose a novel approach for model comparisons when ROC curves show intersections. We investigate in a theoretical framework the relationship between ROC orderings and stochastic dominance and we propose alternative indicators that could substitute the common AUC measure. We also provide empirical evidence based on a real data set.
Making classifier performance comparisons when Receiver Operating Characteristic curves intersect
GIGLIARANO, CHIARA;
2012-01-01
Abstract
The main objective of this paper is to propose a novel approach for model comparisons when ROC curves show intersections. We investigate in a theoretical framework the relationship between ROC orderings and stochastic dominance and we propose alternative indicators that could substitute the common AUC measure. We also provide empirical evidence based on a real data set.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
