Many different statistical methods have been developed for predicting the disease classes of patients. However, in order to have confidence in the results of such methods, their performance needs to be assessed. Different performance measures are reviewed and the circumstances in which they are relevant are described. Subtleties exist which must be taken into account to ensure that the measure chosen matches the objectives. Examples are given showing different interpretations of future diagnostic performance.
No takes yet. Share an insight, caveat, or question.
David J. Hand (2001) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: