Heterogeneous multiple kernel learning for breast cancer outcome evaluation

BMC Bioinformatics

Table 2 Averaged AUC values for determining the optimal σ in the RBF kernel

Datasets	σ = 0.01	σ = 0.1	σ = 1	σ = 10	σ = 100	σ = 1000
GSE32394	0.1589 ± 0.1189	0.1511 ± 0.1511	0.1956 ± 0.1400	0.6667 ± 0.1667	0.9344 ± 0.0456	0.9367 ± 0.0700
GSE59993	0.3455 ± 0.0637	0.3606 ± 0.1239	0.4286 ± 0.0433	0.8287 ± 0.0247	0.6891 ± 0.0412	0.6988 ± 0.0413
GSE1872	0.2697 ± 0.0917	0.2042 ± 0.0686	0.2068 ± 0.0659	0.2432 ± 0.1053	0.2424 ± 0.1061	0.2458 ± 0.1027
GSE76260	0.3823 ± 0.0796	0.4224 ± 0.0464	0.3837 ± 0.0937	0.8270 ± 0.0168	0.8357 ± 0.0213	0.8337 ± 0.0485
GSE59246	0.4550 ± 0.0543	0.4442 ± 0.0785	0.7543 ± 0.0462	0.7539 ± 0.0334	0.7553 ± 0.0111	0.7629 ± 0.0094
BRCA1	0.2565 ± 0.0776	0.2336 ± 0.1205	0.4720 ± 0.1095	0.9918 ± 0.0060	0.9659 ± 0.0303	0.9407 ± 0.0951
BRCA2	0.2316 ± 0.0497	0.2377 ± 0.1074	0.3709 ± 0.1072	1.0000 ± 0.0000	1.0000 ± 0.0000	1.0000 ± 0.0000
BRCA3	0.3410 ± 0.0424	0.3351 ± 0.0335	0.7377 ± 0.1495	1.0000 ± 0.0000	1.0000 ± 0.0000	1.0000 ± 0.0000

ISSN: 1471-2105