Table 3.

The model evaluation for second primary malignancies and recurrent cancer co-discussion.

	Classifier	TP	FP	TN	FN	Sn	Sp	Acc	MCC
SPM+Re	REPTree	161	69	141	47	0.774	0.671	0.722	0.448
	REPTree_F	169	83	127	39	0.813	0.605	0.708	0.426
	LIBSVM	145	63	148	62	0.700	0.701	0.701	0.402
	LIBSVM_F	161	65	145	47	0.774	0.690	0.732	0.466
NSPM+Re	REPTree	1546	365	1416	235	0.868	0.795	0.832	0.665
	REPTree_F	1572	374	1407	209	0.883	0.790	0.836	0.676
	LIBSVM	1477	365	1416	304	0.829	0.795	0.812	0.625
	LIBSVM_F	1569	370	1411	212	0.881	0.792	0.837	0.676
SMP+NRe	REPTree	252	104	229	81	0.757	0.688	0.722	0.446
	REPTree_OP	274	108	225	59	0.823	0.676	0.749	0.504
	REPTree_F	236	98	235	97	0.709	0.706	0.707	0.414
	LIBSVM	235	99	234	98	0.706	0.703	0.704	0.408
	LIBSVM_F	226	63	270	107	0.679	0.811	0.745	0.494
NSPM+NRe	REPTree	1705	504	1473	272	0.862	0.745	0.804	0.612
	REPTree_OP	1739	505	1472	238	0.880	0.745	0.812	0.630
	REPTree_F	1670	465	1512	307	0.845	0.765	0.805	0.611
	LIBSVM	1700	539	1438	277	0.860	0.727	0.794	0.592
	LIBSVM_2F	1705	502	1475	272	0.862	0.746	0.804	0.613

LIBSVM_F indicates SVM model building with feature selection. REPTree_F indicates REPTree model building with feature selection. REPTree_OP indicates REPTree model building by parameters optimization.