“…Bellotti and Crook (2009) use SVM, LR, LDA and kNN on a very large data set (25,000 records) from a financial institution and find that SVM is comparatively successful in classifying credit card debtors who do default; but unlike the other compared models, a large number of support vectors are required to achieve the best performance. Two comparative studies (Zurada, 2007(Zurada, , 2010) use LR, NN, DT, memory-based reasoning (MBR) and an ensemble model using German and SAS-1 data sets. Both found that for some cut-off points and conditions, DTs perform well with respect to classification accuracy and that DTs are attractive tools for decision makers because they can generate easy to interpret if-then rules.…”