Journal of Big Data

Table 7 Tukey HSD test for logistic regression, Naïve Bayes, and MLP, ordered by AUPRC

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

LR			NB			MLP
P	AUPRC	Group	P	AUPRC	Group	P	AUPRC	Group
1700	0.3256	a	2500	0.3893	a	1700	0.2899	a
1600	0.3241	a	2400	0.3879	ab	1900	0.2825	ab
1500	0.3226	a	2300	0.3869	ab	2000	0.2766	abc
1900	0.3220	a	2200	0.3866	abc	1800	0.2763	abc
1800	0.3219	a	1900	0.3861	abc	2100	0.2735	abc
2500	0.3191	a	2100	0.3860	abc	1500	0.2692	abcd
2000	0.3118	ab	2000	0.3860	abc	2200	0.2688	abcd
2200	0.3117	ab	1700	0.3852	abc	1600	0.2685	abcd
2100	0.3108	ab	1800	0.3851	abc	2300	0.2672	abcd
1400	0.3108	ab	1300	0.3842	abcd	2500	0.2659	abcd
2300	0.3106	ab	1600	0.3838	abcd	2400	0.2651	abcd
2400	0.3093	abc	1400	0.3837	abcd	1300	0.2576	bcde
1300	0.3074	abc	1500	0.3833	abcd	1400	0.2517	cde
1200	0.2831	bcd	1200	0.3826	abcd	1100	0.2476	cdef
1100	0.2818	bcd	1000	0.3814	abcd	1200	0.2409	def
1000	0.2793	cd	1100	0.3814	abcd	1000	0.2345	efg
900	0.2759	d	900	0.3791	abcd	900	0.2198	fgh
800	0.2610	de	800	0.3779	abcd	800	0.2090	gh
700	0.2607	de	700	0.3735	bcd	700	0.1962	h
600	0.2355	ef	600	0.3717	cd	600	0.1935	h
500	0.2301	f	500	0.3693	d	500	0.1904	h

Back to article page