The Construction of Support Vector Machine Classifier Using the Firefly Algorithm

Table 5

Performance evaluation for each disease group.

Method used	Sensitivity (%)	Specificity (%)	False negative rate (%)	Accuracy (%)
Method 1: firefly-SVM based OAA-FSVM			2.5	92.50 ± 1.37
() Normal	93.10	90.00
() Inflammation tendon	93.10	96.42
() Calcific tendon	96.67	93.10
() Supraspinatus tear	100.00	100.00
Method 2: original OAA-FSVM trained by LIBSVM			3.33	89.10 ± 2.49
() Normal	83.33	95.56
() Inflammation tendon	86.67	95.56
() Calcific tendon	90.00	96.67
() Supraspinatus tear	100.00	100.00