A Novel Support Vector Machine with Globality-Locality Preserving

Table 4

Classification accuracy for comparative analysis.

Database	SVM	SVM + LDA	MCVSVM	MCLPV_SVM	GLPSVM
Breast	0.8613±0.0103	0.9572±0.0063	0.9596±0.0052	0.9658±0.0045	𝟎.𝟗𝟔𝟖𝟔±𝟎.𝟎𝟎𝟒𝟔
𝐶=1				𝑘=40	𝑘1=40,𝑘2=1
				𝑡=2	𝜎1=𝜎2=2
Heart	0.6912±0.0352	0.8176±0.0205	0.8235±0.0186	0.8294±0.0338	𝟎.𝟖𝟒𝟕𝟏±𝟎.𝟎𝟐𝟒𝟑
𝐶=10				𝑘=50	𝑘1=50,𝑘2=1
				𝑡=2	𝜎1=𝜎2=2
Pima	0.6723±0.0296	0.7689±0.0158	0.7663±0.0144	0.7689±0.1023	𝟎.𝟕𝟖𝟒𝟏±𝟎.𝟎𝟐𝟕𝟕
𝐶=10				𝑘=80	𝑘1=80,𝑘2=1
				𝑡=2	𝜎1=𝜎2=2
NT	0.6923±0.0000	0.8965±0.0157	0.8569±0.0148	0.9496±0.0144	𝟎.𝟗𝟓𝟏𝟐±𝟎.𝟎𝟏𝟕𝟎
𝐶=100				𝑘=10	𝑘1=10,𝑘2=2
				𝑡=0.5	𝜎1=𝜎2=0.5
Wine	0.5714±0.0275	0.9490±0.0228	0.8163±0.0126	0.9694±0.0219	𝟎.𝟗𝟕𝟗𝟔±𝟎.𝟎𝟐𝟎𝟕
𝐶=100				𝑘=20	𝑘1=20,𝑘2=2
				𝑡=2	𝜎1=𝜎2=2
Iris	0.9467±0.0172	0.9713±0.0139	0.9733±0.0147	0.9333±0.0140	𝟎.𝟗𝟕𝟓𝟑±𝟎.𝟎𝟏𝟐𝟒
𝐶=1				𝑘=10	𝑘1=10,𝑘2=2
				𝑡=0.5	𝜎1=𝜎2=0.5
Glass	0.6508±0.0179	0.7996±0.0132	0.7967±0.0098	0.9172±0.2222	𝟎.𝟗𝟒𝟎𝟐±𝟎.𝟎𝟐𝟏𝟗
𝐶=100				𝑘=4	𝑘1=4,𝑘2=4
				𝑡=0.5	𝜎1=𝜎2=0.5