A Semisupervised Cascade Classification Algorithm

Algorithm 1

Cascading NB and C4.5 (NBC4.5).

Algorithm  CascadeGeneralization(L1, L2, Dataset)

begin

  Dataset′  =  ExtendDataset(Dataset, , Dataset)

   = model induced by from Dataset′

  For each new query instance

    = ExtendDataset(Dataset, , )

  Class =

end

where:

   (i) is the base level learning algorithm (NB Classifier in all the training set e.g calculate for each class , and returns

    the class with the max probability)

   (ii) is the meta-level learning algorithm: C4.5

Algorithm  

begin

   = model induced by from

  For each

    For to

        = probability that belongs to according to

        = (, predicted class of : )

  Return

end

where:

   (i) is a learning algorithm, is the number of classes

   (ii) is the dataset from which builds a model, is a dataset to which the model learnt from may be applied