Prediction of Defective Software Modules Using Class Imbalance Learning

Table 7

Comparison on the basis of geometric mean values on eight datasets.

Dataset	SVM	CBNN	NB	RF	LR	-NN	BBN	C4.5	LSTSVM	WLSTSVM
CM1	0.3809	0.6496	0.6039	0.4358	0.4024	0.5125	0.4938	0.3355	0.7088	0.7198
KC1	0.4399	0.6918	0.5389	0.4436	0.5108	0.5685	0.4672	0.4766	0.7039	0.7173
PC1	0.7314	0.6718	0.5673	0.5378	0.4425	0.4803	0.6605	0.4848	0.7247	0.7345
PC3	0.6191	0.6640	0.5086	0.3971	0.3904	0.4801	0.4305	0.4315	0.6138	0.6438
PC4	0.7652	0.7312	0.5807	0.580	0.4838	0.5052	0.5984	0.4899	0.7104	0.7505
MC2	0.6273	0.6036	0.5615	0.6109	0.3834	0.4659	0.5433	0.4270	0.5960	0.6345
KC2	0.7150	0.6993	0.6874	0.6212	0.5964	0.6780	0.6335	0.5562	0.7479	0.7788
KC3	0.5538	0.6138	0.6058	0.4860	0.3407	0.6417	0.5807	0.4302	0.6511	0.6642