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

Table 5

Comparison of the classification results (%) on BUPA liver dataset.

	Method	Sensitivity	Specificity	Accuracy
Linear	FSVM	88.10±0.008	66.42±0.073	72.19±0.057
	SVDD	87.27±0.021	68.05±0.063	72.72±0.042
	FSVM-CIL	88.00±0.004	67.44±0.042	73.19±0.015
	WCS-FSVM	84.00±0.360	67.15±0.068	71.66±0.051
	FSVM-CIPlin	88.00±0.004	67.88±0.063	73.26±0.031
	FSVM-CIPexp	86.00±0.048	69.34±0.072	𝟕𝟑.𝟖𝟎±𝟎.𝟎𝟓𝟒
Gaussian kernel	FSVM	96.00±0.057	66.67±0.026	74.60±0.038
	SVDD	95.43±0.033	71.24±0.050	77.23±0.017
	FSVM-CIL	95.00±0.045	72.59±0.052	78.37±0.050
	WCS-FSVM	90.08±0.070	67.44±0.083	73.73±0.062
	FSVM-CIPlin	94.00±0.049	74.10±0.045	79.46±0.048
	FSVM-CIPexp	94.00±0.049	73.33±0.084	𝟕𝟗.𝟗𝟐±𝟎.𝟎𝟕𝟒