Improving the predictive accuracy of the cross-selling of consumer loans using deep learning networks

Abstract Traditionally most cross-selling models in retail banking use demographics information and interactions with marketing as input to statistical or machine learning algorithms predict whether a customer is willing purchase given financial product not. We overcome such limitation by building several that also years of account transaction data. The objective this study analysis credit card transactions customers, order come up good prediction products. deep-learning algorithm analyze almost 800,000 cards transactions. results show unique data contains valuable on the customers’ consumption behavior it can significantly increase predictive accuracy model. In summary, we develop an auto-encoder extract features from them classifier. demonstrate have power enhances performance model even further.