PMSVM: An Optimized Support Vector Machine Classification Algorithm Based on PCA and Multilevel Grid Search Methods

Table 4

Comparison of the accuracy and efficiency of algorithms on random sampled datasets.

Dataset	Algorithm	Parameters		Time (S)	Accuracy
					Sensitivity	Specificity	Precision	AUC	CVA (%)
R-DATA2	LIBSVM	0.50000	0.1767800	19.610260	0.805	0.857	0.825	0.90640	82.2222
	PMSVM	256.000	0.0039063	5.5637770	0.780	0.857	0.825	0.96925	83.8889
R-DATA4	LIBSVM	1.00000	0.0888388	19.566909	0.842	0.913	0.877	0.94345	87.7778
	PMSVM	2.82840	0.0883880	4.1918490	0.842	0.923	0.889	0.95078	87.7778
R-DATA6	LIBSVM	1.00000	0.0220970	19.573738	0.802	0.909	0.878	0.93628	86.1111
	PMSVM	8.00000	0.0156250	4.5530740	0.802	0.929	0.903	0.94762	87.2222
R-DATA8	LIBSVM	0.17678	1.0000000	19.871964	0.785	0.911	0.873	0.89034	83.8889
	PMSVM	4.00000	0.0312500	5.1134920	0.797	0.861	0.818	0.91580	83.3333
R-DATA10	LIBSVM	0.12500	0.1767800	19.638424	0.750	0.885	0.826	0.89020	84.4444
	PMSVM	0.25000	0.0883880	4.8284440	0.763	0.885	0.829	0.89890	83.8889