New Fuzzy Support Vector Machine for the Class Imbalance Problem in Medical Datasets Classification

Table 3

Comparison of the classification results (%) on heart dataset.

	Method	Sensitivity	Specificity	Accuracy
Linear	FSVM	87.50±0.080	80.77±0.069	82.35±0.069
	SVDD	87.03±0.021	77.69±0.005	80.00±0.051
	FSVM-CIL	85.00±0.046	82.04±0.110	82.35±0.072
	WCS-FSVM	87.30±0.071	81.54±0.089	82.94±0.088
	FSVM-CIPlin	85.00±0.063	82.31±0.083	82.84±0.054
	FSVM-CIPexp	87.50±0.025	82.31±0.083	𝟖𝟑.𝟓𝟑±𝟎.𝟎𝟓𝟓
Gaussian kernel	FSVM	86.70±0.099	82.61±0.087	83.35±0.042
	SVDD	90.35±0.022	80.77±0.034	82.80±0.070
	FSVM-CIL	87.05±0.034	81.54±0.067	82.94±0.044
	WCS-FSVM	91.00±0.076	81.73±0.083	84.12±0.085
	FSVM-CIPlin	90.00±0.045	82.31±0.086	84.12±0.052
	FSVM-CIPexp	86.05±0.023	83.08±0.078	𝟖𝟒.𝟕𝟏±𝟎.𝟎𝟔𝟔