Skip to main content

Table 4 The performance of different kernels (RBF kernel) on PDB1075 data set (leave one out)

From: A sequence-based multiple kernel model for identifying DNA-binding proteins

Kernel type Model ACC (%) SN (%) Spec (%) MCC AUC
\({\mathbf {K}}_{GE}\) SVM 71.6 70.1 73.1 0.432 0.785
\({\mathbf {K}}_{MCD}\) SVM 70.9 68.2 73.5 0.417 0.761
\({\mathbf {K}}_{NMBAC}\) SVM 71.1 73.3 69.1 0.424 0.771
\({\mathbf {K}}_{PSSM-AB}\) SVM 76.9 84.4 69.8 0.547 0.839
\({\mathbf {K}}_{PSSM-DWT}\) SVM 76.0 79.2 72.9 0.522 0.837
\({\mathbf {K}}_{PsePSSM}\) SVM 78.5 82.5 74.7 0.573 0.857
Mean weighted kernels SVM 83.1 84.6 81.8 0.664 0.913
MKL-CKA SVM 84.2 85.9 82.6 0.684 0.914
  1. The bold font indicates the largest value in the column