Additional File 2.

SVM parameter optimization Training sets of different sequence window datasets were trained under 10-fold cross-validation using various combinations of C and γ values. The optimal C and γ values for each training set are used to train the final SVM classifier. Optimal values are indicated below the corresponding chart.

Format: PDF Size: 611KB Download file

This file can be viewed with: Adobe Acrobat Reader

Wee et al. BMC Genomics 2011 12(Suppl 3):S11   doi:10.1186/1471-2164-12-S3-S11