The problem in the employed evaluation method is the
availability of sufficient data set to be divided as training data
and test data. This research employed the stratified tenfold
cross validation for the evaluation to overcome the limitation
of data availability [10]. To implement this, the data were
divided into ten parts randomly in nearly the same size. Each
part was held out in turn and the nine-tenths were trained using
classification algorithms. Afterward, the error rates were
calculated. Thereby, the algorithm was executed ten times on
different training sets. At last, the overall error rate was
resulted from the average of ten error estimates.