Prediction of Defective Software Modules Using Class Imbalance Learning

Table 4

Comparison on the basis of specificity values on eight datasets.

Dataset	SVM	CBNN	NB	RF	LR	-NN	BBN	C4.5	LSTSVM	WLSTSVM
CM1	0.9603	0.7148	0.8286	0.5883	0.5772	0.6836	0.5322	0.4274	0.6612	0.6684
KC1	0.9745	0.7002	0.9405	0.6224	0.5818	0.7082	0.6501	0.5301	0.6435	0.8040
PC1	0.8078	0.8342	0.9028	0.6917	0.6132	0.7773	0.7804	0.6022	0.9750	0.7946
PC3	0.6004	0.7282	0.9066	0.7009	0.5569	0.5491	0.5845	0.4867	0.5742	0.6981
PC4	0.8027	0.8136	0.8750	0.6173	0.4936	0.5518	0.6173	0.4682	0.6641	0.7180
MC2	0.7556	0.4664	0.9108	0.6702	0.4080	0.5422	0.6867	0.3288	0.4655	0.5118
KC2	0.7385	0.7684	0.8143	0.6875	0.5934	0.6702	0.6224	0.5304	0.7391	0.7773
KC3	0.9164	0.7429	0.7976	0.5402	0.4127	0.6936	0.5873	0.4566	0.5506	0.5882