Journal of Big Data

Table 5 Average classifier AUPRC performance results with respect to number of positive-labeled instances P

From: Synthesizing class labels for highly imbalanced credit card fraud detection data

AUPRC
P	DT	RF	ET	LR	NB	MLP
500	0.2935	0.3394	0.3357	0.2301	0.3693	0.1904
600	0.3008	0.3404	0.3491	0.2355	0.3717	0.1935
700	0.3029	0.3493	0.3569	0.2607	0.3735	0.1962
800	0.3111	0.3470	0.3574	0.2610	0.3779	0.2090
900	0.3146	0.3607	0.3649	0.2759	0.3791	0.2198
1000	0.3181	0.3529	0.3670	0.2793	0.3814	0.2345
1100	0.3292	0.3523	0.3675	0.2818	0.3814	0.2476
1200	0.3363	0.3487	0.3613	0.2831	0.3826	0.2409
1300	0.3542	0.3602	0.3795	0.3074	0.3842	0.2576
1400	0.3565	0.3608	0.3806	0.3108	0.3837	0.2517
1500	0.3714	0.3596	0.3842	0.3226	0.3833	0.2692
1600	0.3751	0.3615	0.3832	0.3241	0.3838	0.2685
1700	0.3774	0.3597	0.3859	0.3256	0.3852	0.2899
1800	0.3801	0.3623	0.3834	0.3219	0.3851	0.2763
1900	0.3818	0.3547	0.3825	0.3220	0.3861	0.2825
2000	0.3764	0.3491	0.3741	0.3118	0.3860	0.2766
2100	0.3737	0.3393	0.3712	0.3108	0.3860	0.2735
2200	0.3767	0.3328	0.3683	0.3117	0.3866	0.2688
2300	0.3758	0.3221	0.3650	0.3106	0.3869	0.2672
2400	0.3763	0.3125	0.3592	0.3093	0.3879	0.2651
2500	0.3813	0.3009	0.3616	0.3191	0.3893	0.2659

Back to article page