Research on Cross-Company Defect Prediction Method to Improve Software Security

Table 8

Sample weight setting performance analysis results.

No.	Source	Target	AUC-MM	AUC-MMWS
1	CM1	Ant-1.7	0.746	0.69
2	MW1	Ant-1.7	0.707	0.74
3	PC1	Ant-1.7	0.641	0.708
4	PC3	Ant-1.7	0.748	0.743
5	PC4	Ant-1.7	0.734	0.715
6	CM1	Camel-1.6	0.549	0.558
7	MW1	Camel-1.6	0.536	0.629
8	PC1	Camel-1.6	0.55	0.617
9	PC3	Camel-1.6	0.549	0.622
10	PC4	Camel-1.6	0.589	0.544
11	CM1	Ivy-2.0	0.761	0.708
12	MW1	Ivy-2.0	0.71	0.772
13	PC1	Ivy-2.0	0.603	0.769
14	PC3	Ivy-2.0	0.756	0.723
15	PC4	Ivy-2.0	0.729	0.687
16	CM1	Jedit-4.3	0.689	0.711
17	MW1	Jedit-4.3	0.693	0.707
18	PC1	Jedit-4.3	0.674	0.743
19	PC3	Jedit-4.3	0.678	0.67
20	PC4	Jedit-4.3	0.7	0.703
21	CM1	Synapse-1.2	0.635	0.648
22	MW1	Synapse-1.2	0.676	0.667
23	PC1	Synapse-1.2	0.656	0.685
24	PC3	Synapse-1.2	0.665	0.632
25	PC4	Synapse-1.2	0.65	0.711
26	CM1	Velocity-1.6	0.68	0.663
27	MW1	Velocity-1.6	0.593	0.68
28	PC1	Velocity-1.6	0.599	0.679
29	PC3	Velocity-1.6	0.596	0.686
30	PC4	Velocity-1.6	0.579	0.67
31	CM1	Xalan-2.4	0.602	0.684
32	MW1	Xalan-2.4	0.611	0.726
33	PC1	Xalan-2.4	0.6	0.717
34	PC3	Xalan-2.4	0.663	0.716
35	PC4	Xalan-2.4	0.607	0.703
Average			0.65	0.686