Skip to main content

Advertisement

Table 4 Comparison of kernels in terms of maximum Az value of calcification dataset

From: AdaBoost-based multiple SVM-RFE for classification of mammograms in DDSM

kernel type MGH WU WUFSM SHH
  8 22 8 22 8 22 8 22
linear 0.72686 0.72625 0.89981 0.90870 0.74046 0.77509 0.89603 0.92705
RBF 0.91042 0.76826 0.99192 0.88155 0.93625 0.89079 0.96280 0.94826
C 1 10 1 5 10 20 10 10
γ 1.5 0.1 1 0.05 0.4 0.05 0.15 0.05
  1. Same tradeoff parameter value C is used for both linear and RBF kernels.