Weighted Classification Error Rate Estimator for the Euclidean Distance Classifier
Mindaugas Gvardinskas · Communications in computer and information science · 2015
Error counting estimators are among the best known and most widely used error estimation techniques. Perhaps the best known subcategory of error-counting estimators are k -fold cross-validation methods. Like most other error estimation techniques, cross-validation methods are biased. One way to correct this bias is to use a weighted average of cross-validation and resubstitution estimators. In this paper we propose a new weighted error-counting classification error rate estimator designed specially for the Euclidean distance classifier. Experiments with real world and synthetic data sets show that resubstitution, repeated 2-fold cross-validation, leave-one-out, basic bootstrap and D-method are outperformed by the proposed weighted error rate estimator (in terms of root-mean-square error).