Skip to main content

Table 1 Optimized parameter values for trained classifiers.

From: Pripper: prediction of caspase cleavage sites from whole proteomes

Classifier Parameter Value
SVM SVM method NU-SVC
  Kernel type RBF
  Kernel function parameter γ 2-5.5
  Error parameter ν 0.536
  Stopping criterion ε 0.00001
  Sequence length 10
   6 before cut site
   4 after cut site
Random forest Maximum depth unlimited
  Number of features 4
  Number of trees 143
  Sequence length 24
   12 before cut site
   12 after cut site
J48 Confidence factor 0.285
  Minimum number of objects 5
  Number of folds 3
  Binary splits true
  Reduced error pruning false
  Subtree raising true
  Unpruned false
  Use Laplace false
  Sequence length 6
   4 before cut site
   2 after cut site