Skip to main content

Table 4 Difference between performance measures with and without filtering over all datasets

From: Improving binary classification using filtering based on k-NN proximity graphs

ClassifierDT (%)LR (%)NB (%)SVM (%)NN (%)RF (%)DES-LA (%)
ACC2.970.381.650.110.070.350.57
Sens6.140.853.070.261.412.591.12
Spec7.120.747.92− 0.15− 2.985.42− 1.08
AUC− 1.534.533.76− 1.54− 2.564.230.17
BS− 1.321.610.050.110.811.00− 0.13