Investigating Tree Family Machine Learning Techniques for a Predictive System to Unveil Software Defects

Table 6

True-positive rate analysis by each TF-ML technique on individual dataset.

Technique	AR1	AR3	CM1	KC2	KC3	MW1	PC1	PC2	PC3	PC4
CDT	0.926	0.873	0.894	0.83	0.82	0.923	0.935	0.996	0.892	0.892
CS-Forest	0.851	0.841	0.825	0.791	0.814	0.883	0.912	0.995	0.848	0.889
DS	0.909	0.905	0.902	0.797	0.82	0.911	0.931	0.996	0.898	0.878
Forest-PA	0.917	0.889	0.9	0.835	0.799	0.921	0.935	0.996	0.897	0.901
HT	0.926	0.841	0.902	0.833	0.809	0.923	0.931	0.996	0.898	0.881
J48	0.901	0.873	0.88	0.814	0.794	0.921	0.933	0.996	0.889	0.894
LMT	0.917	0.873	0.892	0.843	0.794	0.931	0.924	0.996	0.891	0.903
RF	0.901	0.921	0.892	0.833	0.814	0.921	0.937	0.996	0.901	0.907
RT	0.893	0.873	0.833	0.808	0.706	0.866	0.911	0.992	0.855	0.867
REP-T	0.926	0.873	0.892	0.816	0.794	0.916	0.936	0.996	0.896	0.897