Table 1.

Prediction Results of Dataset-2 Using Various Linear Classifiers

		Number of Z curve variables
		69				189				252
Length	Classifiers				Std				Std				Std
192 bp	FDA	96.36	96.74	96.55	/	96.09	96.99	96.54	/	88.93	85.41	87.17	0.47
	PLS	98.52	95.49	97.01	0.89	98.40	96.23	97.31	0.90	98.64	96.39	97.52	0.90
	LS	98.23	95.16	96.70	0.89	98.32	95.65	96.98	0.90	91.43	89.38	90.4	1.19
	RR	98.23	95.12	96.68	0.89	98.52	95.57	97.05	0.90	98.56	96.02	97.29	0.90
42 bp	FDA	90.84	80.74	80.79		80.18	81.50	80.84		77.31	73.39	75.35	0.56
	PLS	87.61	78.75	83.18	0.70	87.82	81.95	84.88	0.75	88.44	82.46	85.45	0.78
	LS	87.20	76.38	81.79	0.71	85.65	78.54	82.10	0.77	82.67	77.00	79.83	1.09
	RR	89.78	74.22	82.00	0.69	89.78	74.53	82.15	0.72	89.78	75.66	82.72	0.74
30 bp	FDA	82.83	68.17	75.50		77.26	71.10	74.18		73.74	67.00	70.37	0.41
	PLS	82.83	73.16	77.99	0.65	81.07	74.33	77.70	0.84	81.36	76.38	78.87	0.79
	LS	81.95	69.93	75.94	0.68	71.98	69.64	70.81	0.84	71.40	64.95	68.17	1.19
	RR	87.81	60.55	74.18	0.59	82.53	62.90	72.72	0.70	78.43	65.53	71.98	0.75
21 bp	FDA	71.29	71.78	71.53		68.81	69.80	69.30		63.86	62.87	63.37	0.28
	PLS	74.26	79.70	76.98	0.64	76.73	79.70	77.22	0.82	78.71	79.70	79.21	0.79
	LS	73.28	71.29	72.28	0.71	71.78	69.80	70.79	1.04	70.30	66.34	68.32	1.41
	RR	80.20	55.94	68.07	0.57	80.69	64.36	72.52	0.75	75.74	71.78	73.76	0.92

The average accuracies of PLS models, which were the best ones among the algorithms evaluated here, are shown in boldface. The results of FDA calculated by Gao and Zhang (2004) are shown in italics. Std, standard deviations of prediction results; FDA, Fisher discriminant analysis; LS, least squares; PLS, partial least squares; RR, ridge regression.