L2-norm multiple kernel learning and its application to biomedical data fusion

BMC Bioinformatics

Table 10 Results of experiment 4 data set III: classification of PUL patients using multiple kernels derived from clinical data

Classifier	Mean - error of AUC	Std. - error of AUC	pvalue
Weighted LSSVM L ₂ MKL	0.1165	0.0100	-
Weighted LSSVM L ₁ MKL	0.1243	0.0171	0.0519
Weighted LSSVM L_∞ (0.5) MKL	0.1290	0.0206	0.0169
Weighted SVM L₂ MKL	0.1499	0.0248	4.79 · 10^-5
Weighted SVM L_∞ MKL	0.1552	0.0210	1.02 · 10-6
Weighted SVM L_∞ (0.5)	0.1551	0.0153	3.87 · 10^-6
Weighted SVM L₁ MKL	0.1594	0.0162	2.29 · 10^-9
Weighted LSSVM L_∞ MKL	0.1651	0.0174	4.41 · 10^-10

Results of experiment 4 data set II: classification of PUL patients using multiple kernels derived from clinical data. The classifier with the best performance is shown in bold. The p-values are compared with the best performance using a paired t-test. The performance of classifiers is sorted from high to low according to the p-values.

ISSN: 1471-2105