Data mining feature selection for credit scoringmodels

The features used may have an important effect on the performance of credit scoring models. The process of choosingthe best set of features for credit scoring models is usually unsystematic and dominated by somewhat arbitrary trial.This paper presents an empirical study of four machine learning feature selection methods. These methods provide anautomatic data mining technique for reducing the feature space. The study illustrates how four feature selectionmethods—‘ReliefF’, ‘Correlation-based’, ‘Consistency-based’ and ‘Wrapper’ algorithms help to improve three aspectsof the performance of scoring models: model simplicity, model speed and model accuracy. The experiments areconducted on real data sets using four classification algorithms—‘model tree (M5)’, ‘neural network (multi-layerperceptron with back-propagation)’, ‘logistic regression’, and ‘k-nearest-neighbours’...