Credit Card Fraud Detection Using LSTM Algorithm

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