EBOC: Ensemble-Based Ordinal Classification in Transportation

Algorithm 1

The pseudocode of the proposed EBOC approach.

Algorithm EBOC: Ensemble-Based Ordinal Classification

Inputs:

 D:  the ordinal dataset

 m: the number of instances in the dataset D

 X: input feature space, an input vector xϵX

 Y: class attribute, an ordinal class label , with an order

 k: the number of class labels

 n: ensemble size

Outputs:

 M: an ensemble classification model

 M(x): class label of a new sample x

Begin:

 for  i = 1 to  n  do

  if (bagging)

    = bootstrap samples from D

  else if (boosting)

    =samples from D according to their weights

  // Construction of binary training sets,

  for  j = 1 to  k-1 do

   for  s = 1 to  m  do

    if ( <= )

     Add(,0)

    else

     Add(,1)

   end for

  end for

  // Construction of binary classifiers, BC_ij

  for  j = 1 to  k-1  do

   BC_ij=ClassificationAlgorithm()

  end for

  if (boosting)

   update weight values

 end for

 // Classification of a new sample x

 for  i = 1  to  n  do

  // Construction of ordinal classification models,

  P() = 1 − P(y > )

  for  j = 2  to  k-1  do

   P() = P(y > ) − P(y > )

  end for

  P() = P(y > )

   = max(P)

 end for

 // Majority voting

 if (bagging)

 else if (boosting)

End Algorithm