Table 3.

Number of datasets correctly predicted for both the simple and complex networks combined by case size (bolded values indicated highest score in each case size)

Learning method	300 cases	600 cases	1200 cases	2400 cases	4800 cases	9600 cases	Total
DCLα=54	559	685	815	945	1190	1463	5657
DCLα=15	563	719	825	934	1122	1393	5556
DCLα=9	520	684	796	893	1066	1349	5308
DCLα=108	396	531	703	874	1083	1339	4926
DCLα=1	347	495	708	785	893	1165	4393
FGSsmp2 stp2	235	363	622	880	1000	1176	4276
FGSsmp2 stp1	234	296	586	850	970	1139	4075
FGSsmp1 stp2	232	315	548	857	976	1134	4062
CPCβ=0.1	159	326	556	773	998	1154	3966
FGSsmp1 stp1	230	269	517	829	956	1095	3896
CPCβ=0.2	200	383	541	758	930	1032	3844
CPCβ=0.05	113	265	491	718	982	1167	3736
CPCβ=0.01	54	169	366	608	856	1109	3162
PCβ=0.01	209	273	225	161	132	166	1166
PCβ=0.1	141	155	154	146	158	215	969
PCβ=0.05	159	158	152	123	160	200	952
PCβ=0.2	140	86	116	142	146	199	829
FCIβ=0.2	3	3	5	13	16	24	64
FCIβ=0.1	3	1	1	3	5	15	28
FCIβ=0.05	2	1	0	2	4	4	13
FCIβ=0.01	0	1	0	0	1	0	2