Table 1.
An overview of the typical technologies of CRR.
Fig 1.
The flowchart of the study.
Fig 2.
Compositions of data acquisition system for CRR.
Table 2.
Feature attributes of the C-R dataset after feature selection.
Fig 3.
Two-class MEB-SVM classifier.
Fig 4.
Mapping processing from input space to MED feature space (n = 2).
Fig 5.
Euclidean distance in the constructed balls.
Table 3.
Details of the datasets from UCI repository used in the experiments.
Table 4.
Accuracies of experiments comparing with the referenced algorithms.
Table 5.
Critical values for the two-tailed Nemenyi test after the Friedman test.
Table 6.
Critical values for the two-tailed Bonferroni-Dunn test after the Friedman test.
Table 7.
Comparison of AUC between eight algorithms.
Fig 6.
Bonferroni-Dunn test graphic.
Table 8.
Test accuracy (in %) for single feature variable subsets with MEB-SVM.
Fig 7.
Predictive accuracy values of MFS+MEB-SVM, MEB-SVM, PSO + SVM and SVM.