Table 8.

Mean logarithmic losses of different algorithms based on 10-fold cross-validation done twice. The bottom row gives the average logarithmic losses. Best results are underlined.

Data Set	ISMB	ISMB-MS	NISMB	NB	DT	LR	NN	kNN	LBR	AB
australian	0.3390	0.3456	0.3417	0.4476	0.4091	0.7136	0.4263	0.8627	0.4482	0.3850
breast-cancer	0.1068	0.1138	0.1083	0.2497	0.2955	0.1485	0.1205	0.2138	0.2497	0.2772
cleveland	0.3925	0.4067	0.4021	0.4491	1.3001	0.6500	0.4584	0.8625	0.4491	0.7044
corral	0.1018	0.1101	0.0989	0.3326	0.1475	0.2753	0.1542	0.0175	0.3130	0.1280
crx	0.3451	0.3564	0.3525	0.4113	0.3783	0.9377	0.4678	0.8747	0.4018	0.3845
diabetes	0.4601	0.4606	0.4604	0.4809	0.5497	0.4588	0.6039	0.5028	0.4826	0.4478
flare	0.4282	0.4294	0.4314	0.5904	0.4879	0.4042	0.5333	0.5858	0.5182	0.4032
german	0.5331	0.5413	0.5377	0.5213	1.4604	0.5229	0.5801	1.5415	0.5221	0.7981
glass2	0.4238	0.4302	0.4246	0.4532	0.8498	0.4154	0.8853	0.4562	0.4447	0.4530
glass	0.7112	0.7239	0.7113	0.7697	2.3005	4.0749	1.3612	0.8685	0.7264	0.9452
heart	0.3996	0.4069	0.3973	0.4560	0.6920	0.3907	0.6109	0.8483	0.4560	0.3788
hepatitis	0.2396	0.2517	0.2583	0.4247	0.6122	17.7871	0.3562	0.6253	0.4272	0.2548
iris	0.1560	0.1909	0.1620	0.1621	0.5287	0.7579	0.5770	0.2240	0.1621	0.1712
lymphography	0.4100	0.4289	0.4430	0.4282	2.9112	21.6371	0.5765	0.7272	0.4409	0.7422
pima	0.4647	0.4657	0.4657	0.4793	0.5268	0.4572	0.5873	0.5114	0.4774	0.4880
postoperative	0.7381	0.7776	0.7287	0.7953	1.1395	2.8236	1.3339	1.9418	0.7953	0.9453
sonar	0.3573	0.3726	0.3743	0.4573	1.2814	0.5762	0.4170	0.5728	0.4554	0.4344
vehicle	0.5863	0.5900	0.5866	1.8645	2.3842	3.9997	1.0134	1.2590	0.7815	1.0042
vote	0.1393	0.1635	0.1588	0.6804	0.3028	5.5427	0.3171	0.2782	0.5629	0.1562
wine	0.0418	0.0402	0.0367	0.0303	0.8270	0.9593	0.1032	0.0409	0.0303	0.0531
zoo	0.1297	0.1202	0.1268	0.1474	1.1102	0.5325	0.0596	0.1595	0.1474	0.1130
average	0.3573	0.3679	0.3622	0.5063	0.9759	3.0507	0.5497	0.6654	0.4425	0.4604