Mining causal relationships among clinical variables for cancer diagnosis based on Bayesian analysis

Table 2 Experimental results of zero-one loss and standard deviation

Data set	NB	TAN	KDB	FKBN
BREAST	0.191 ±0.002	0.166 ±0.001	0.164 ±0.001	0.162 ±0.001
FEMGEN	0.203 ±0.002	0.149 ±0.002	0.131 ±0.001	0.128 ±0.001
LYMYLEUK	0.297 ±0.002	0.269 ±0.001	0.261 ±0.001	0.257 ±0.001
MALEGEN	0.220 ±0.002	0.163 ±0.002	0.169 ±0.002	0.160 ±0.002
COLRECT	0.197 ±0.002	0.180 ±0.001	0.177 ±0.001	0.175 ±0.001
URINARY	0.205 ±0.003	0.180 ±0.002	0.172 ±0.002	0.169 ±0.002
ALL-AML_Leukemia	0.290 ±0.405	0.289 ±0.421	0.290 ±0.435	0.289 ±0.429
DLBCL-Harvard	0.247 ±0.135	0.247 ±0.147	0.246 ±0.132	0.246 ±0.129
DLBCL-Stanford	0.489 ±0.289	0.489 ±0.309	0.489 ±0.317	0.489 ±0.277
Lung Cancer-Michigan	0.104 ±0.102	0.103 ±0.089	0.104 ±0.110	0.104 ±0.107
MLL_Leukemia	0.649 ±0.257	0.648 ±0.261	0.649 ±0.248	0.646 ±0.249
ProstateCancer	0.490 ±0.113	0.490 ±0.110	0.491 ±0.107	0.489 ±0.116

The first six data sets are selected from the SEER database, the next six data sets are selected from Kent Ridge Bio-medical (KRBM) repository.

ISSN: 1756-0381