Search

Abstract With the rapid growth of consumer credit and the huge amount of financial data developing effective credit scoring models is very crucial. Researchers have developed complex credit scoring models using statistical and artificial intelligence (AI) techniques to help banks and financial institutions to support their financial decisions. Neural networks are considered as a mostly wide used technique in finance and business applications. Thus, the main aim of this paper is to help bank management in scoring credit card clients using machine learning by modelling and predicting the consumer behaviour with respect to two aspects: the probability of single and consecutive missed payments for credit card customers. The proposed model is based on the bidirectional Long-Short Term Memory (LSTM) model to give the probability of a missed payment during the next month for each customer. The model was trained on a real credit card dataset and the customer behavioural scores are analysed using classical measures such as accuracy, Area Under the Curve, Brier score, Kolmogorov–Smirnov test, and H-measure. Calibration analysis of the LSTM model scores showed that they can be considered as probabilities of missed payments. The LSTM model was compared to four traditional machine learning algorithms: support vector machine, random forest, multi-layer perceptron neural network, and logistic regression. Experimental results show that, compared with traditional methods, the consumer credit scoring method based on the LSTM neural network has significantly improved consumer credit scoring.

Abstract One of the ways of increasing recognition ability in classification problem is removing outlier entries as well as redundant and unnecessary features from training set. Filtering and feature selection can have large impact on classifier accuracy and area under the curve (AUC), as noisy data can confuse classifier and lead it to catch wrong patterns in training data. The common approach in data filtering is using proximity graphs. However, the problem of the optimal filtering parameters selection is still insufficiently researched. In this paper filtering procedure based on k-nearest neighbours proximity graph was used. Filtering parameters selection was adopted as the solution of outlier minimization problem: k-NN proximity graph, power of distance and threshold parameters are selected in order to minimize outlier percentage in training data. Then performance of six commonly used classifiers (Logistic Regression, Naïve Bayes, Neural Network, Random Forest, Support Vector Machine and Decision Tree) and one heterogeneous classifiers combiner (DES-LA) are compared with and without filtering. Dynamic ensemble selection (DES) systems work by estimating the level of competence of each classifier from a pool of classifiers. Only the most competent ones are selected to classify a given test sample. This is achieved by defining a criterion to measure the level of competence of base classifiers, such as, its accuracy in local regions of the feature space around the query instance. In our case the combiner is based on the local accuracy of single classifiers and its output is a linear combination of single classifiers ranking. As results of filtering, accuracy of DES-LA combiner shows big increase for low-accuracy datasets. But filtering doesn’t have sufficient impact on DES-LA performance while working with high-accuracy datasets. The results are discussed, and classifiers, which performance was highly affected by pre-processing filtering step, are defined. The main contribution of the paper is introducing modifications to the DES-LA combiner, as well as comparative analysis of filtering impact on the classifiers of various type. Testing the filtering algorithm on real case dataset (Taiwan default credit card dataset) confirmed the efficiency of automatic filtering approach.