Skip to main content

Advertisement

Table 8 Comparison of methods by maximum Az value using 22 features (Calcification)

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

  T MGH WU WFUSM SHH
SVM   0.77497 0.91710 0.89738 0.94945
SVM-RFE   0.77497 0.93436 0.89859 0.95332
ENSEMBLE   0.68951 0.76647 0.72650 0.85677
JOIN (1)   0.75259 0.92326 0.81433 0.91352
JOIN (2)   0.72296 0.82307 0.72987 0.80400
JOIN (3)   0.70815 0.76647 0.70059 0.67598
JOIN (4)   0.58656 0.69779 0.65667 0.55964
JOIN (5)   0.53520 0.63858 0.65667 0.51203
MSVM-RFE (bootstrap) 5 0.77497 0.91710 0.89988 0.95379
  10 0.77826 0.91710 0.89786 0.95330
  15 0.77497 0.92193 0.89738 0.95250
  20 0.77497 0.93305 0.90507 0.95267
MSVM-RFE (boost) 5 0.77727 0.92097 0.89848 0.94945
  10 0.77497 0.93063 0.90108 0.95292
  15 0.77497 0.92352 0.90133 0.95136
  20 0.77497 0.92105 0.89957 0.95256
  1. Numbers in parenthesis stands for cutoff value for JOIN method.