Table 2.

Cross-validation stepwise.

	Training sample size (%)	TP	TN	FP	FN	Sensitivity	Specificity	Accuracy
Binary logistic regression	0.4	78024	9594	4406	9976	0.8866	0.6853	0.8590
	0.6	50115	6878	2122	7885	0.8641	0.7642	0.8506
	0.8	24871	3942	1058	4129	0.8576	0.7884	0.8474
SVM	0.4	81876	7211	6789	6124	0.9304	0.5151	0.8734
	0.6	53604	5122	3878	4396	0.9242	0.5691	0.8765
	0.8	26802	2941	2059	2198	0.9242	0.5882	0.8748
LDA	0.4	79379	10253	3747	8621	0.9020	0.7324	0.8787
	0.6	52929	6682	2318	5071	0.9126	0.7424	0.8897
	0.8	26691	3614	1386	2309	0.9204	0.7228	0.8913
QDA	0.4	83320	3909	10091	4680	0.9468	0.2792	0.8552
	0.6	52004	5052	3948	5996	0.8966	0.5613	0.8516
	0.8	25690	3219	1781	3310	0.8859	0.6438	0.8503
Neural Networks	0.4	85758	7750	6694	2242	0.9745	0.5366	0.9128
	0.6	56802	5317	3683	1198	0.9793	0.5908	0.9271
	0.8	28527	2891	2541	473	0.9837	0.5322	0.9125
Classification Trees	0.4	81399	3459	10541	6601	0.9250	0.2471	0.8319
	0.6	53351	2592	6408	4649	0.9198	0.2880	0.8350
	0.8	26680	1388	3612	2320	0.9200	0.2776	0.8255
Boosting Trees	0.4	80072	3654	7799	1401	0.9828	0.3190	0.9010
	0.6	52540	2480	5605	849	0.9841	0.3067	0.8950
	0.8	26102	1408	5322	455	0.9829	0.2092	0.8264