Data-driven multinomial random forest: a new random forest variant with strong consistency

Journal of Big Data

Table 5 Standard deviation of different RFs on classification and regression datasets

Task	Datasets	DMRF	MRF(SE)	MRF(b)	BRF(SE)	BRF(b)	Denil14(SE)	Denil14(b)	BreimanRF
Classification	Blogger	0.9189	1.4757	2.0789	2.3593	2.0439	1.3984	2.0138	2.1832
	Algerian Forest Fires	0.2907	0.4842	0.7431	0.5495	0.4650	0.3465	0.8440	0.5922
	Wdbc	0.2153	0.3031	0.3751	0.4283	0.3108	0.1860	0.2476	0.4878
	Tic-tac-toe	0.1494	0.1980	0.2642	0.1492	0.3619	0.1690	0.2748	0.2433
	Winequality (Red)	0.3718	0.4873	0.6304	0.5572	0.4748	0.4111	0.5347	0.4871
	Wireless	0.1054	0.0919	0.1061	0.0422	0.1752	0.0577	0.1414	0.1006
	Winequality (white)	0.3107	0.2374	0.2595	0.3072	0.2987	0.1558	0.2839	0.2856
	Connect-4	0.0629	0.0349	0.0627	0.0594	0.0591	0.0262	0.0745	0.0556
Regression	ALE	0.002513	0.002303	0.002802	0.002485	0.002053	0.001867	0.003254	0.002637
	Servo	0.0477	0.0158	0.0200	0.0568	0.0846	0.0201	0.0790	0.0601
	Las Vegas Strip	0.0094	0.0038	0.0162	0.0021	0.3846	0.0011	0.0333	0.0173
	Qsar fish toxicity	0.0105	0.0117	0.0081	0.0116	0.0205	0.0087	0.0131	0.0247
	Flare	0.003023	0.004029	0.011366	0.001570	0.004659	0.000626	0.0101	0.005616
	Winequality (white)	0.003234	0.002213	0.003777	0.001211	0.002729	0.001602	0.005282	0.003536
	Insurance	1.91E-04	1.64E-04	2.27E-04	9.17E-06	1.18E-04	7.57E-06	4.22E-04	1.85E-04
	Combined	0.0738	0.0353	0.0595	0.4213	0.0636	0.1011	0.0709	0.0689