Generalized Resubstitution for Classification Error Estimation

Parisa Ghane, Ulisses Braga-Neto

We propose the family of generalized resubstitution classifier error estimators based on empirical measures. These error estimators are computationally efficient and do not require…