Table 4. Classification of BS2 and WTC-11 cells.

Results of algorithms with the best performances.

	TPR BS2 (%)	TPR WTC-11 (%)	Accuracy (%)	F1 score (%)	MCC (%)
Decision trees	94	98	96	96	92
Naïve Bayes with normal kernel	93	94	93	93	87
SVM with cubic kernel, C = 27	91	95	93	92	86
SVM with quadratic kernel, C = 27	89	94	92	91	84
k-NN with Mahalanobis metric and equal weighting, k = 1	91	88	89	88	79
k-NN with Mahalanobis metric and inverse weighting, k = 1	91	88	89	88	79
k-NN with Mahalanobis metric and squared inverse weighting, k = 1	91	88	89	88	79
SVM with RBF kernel, C = 999	78	95	88	85	76
Naïve Bayes with triangle kernel	75	97	87	84	75
Quadratic discriminant analysis	79	93	87	84	75
k-NN with Cosine metric and squared inverse weighting, k = 5	86	87	87	86	75
k-NN with Cosine metric and equal weighting, k = 1	87	86	86	85	74
k-NN with Cosine metric and inverse weighting, k = 1	87	86	86	85	74
Naïve Bayes with Epanechnikov kernel	72	97	86	82	73
Naïve Bayes with box kernel	70	98	85	81	72
k-NN with City block metric and equal weighting, k = 1	85	85	85	84	72
k-NN with City block metric and inverse weighting, k = 1	85	85	85	84	72
k-NN with City block metric and squared inverse weighting, k = 1	85	85	85	84	72