A Comparative Study: Globality versus Locality for Graph Construction in Discriminant Analysis

Table 1

Accuracy% ± standard deviation (optimal reduced dimensions), where the reduced highest dimension of MDA is number of class subtract one, here equals one.

Datasets (number of sample, number of attr.)	GmLcDA	LmGcDA	LmLcDA (MFA)	LmLcDA (LSDA)	GmGcDA (MDA)	1NN
Crx	70.63 ± 1.22	66.42 ± 5.67	69.76 ± 0.79	68.71 ± 1.74	59.58 ± 7.47	67.29 ± 1.33
(666, 6)	(2)	(6)	(6)	(6)
Sonar	85.15 ± 2.63	62.68 ± 0.65	83.88 ± 1.66	85.63 ± 3.03	69.61 ± 5.81	83.11 ± 3.08
(208, 60)	(42)	(10)	(24)	(34)
Spectf	77.97 ± 2.88	67.52 ± 3.92	76.47 ± 1.77	75.94 ± 1.88	68.72 ± 5.13	71.13 ± 3.90
(267, 44)	( 𝟔 )	(10)	(24)	(4)
Water	95.44 ± 2.06	92.54 ± 0.75	94.91 ± 3.83	92.11 ± 4.40	90.88 ± 4.81	87.90 ± 3.55
(116, 38)	(2)	(4)	(8)	(2)
Wdbc	96.23 ± 1.28	80.51 ± 2.58	95.53 ± 0.86	95.78 ± 0.79	94.97 ± 1.38	91.69 ± 1.13
(569, 30)	( 𝟔 )	(10)	(30)	(10)