Profit-sensitive machine learning classification with explanations in credit risk: The case of small businesses in peer-to-peer lending

We propose a comprehensive profit-sensitive approach for credit risk modeling in P2P lending for small businesses, one of the most financially complex segments. We go beyond traditional and cost-sensitive approaches by including the financial costs and incomes through profits and introducing the profit information at three points of the modeling process: the estimation of the learning function of the classification algorithm (XGBoost in our case), the hyperparameter optimization, and the decision function. The profit-sensitive approaches achieve a higher level of profitability than the profit-insensitive approach in the small business case analyzed by granting mostly lower-risk, lower-amount loans. Explainability tools help us to discover the key features of such loans. Our proposal can be extended to other loan markets or other classification problems as long as the cells of the misclassification matrix have an economic value. •Profit-sensitive approach uses profit info both directly and indirectly.•Methods evaluated in small business, a high-risk segment.•Profit info improves classifier learning and financial performance.•Our approach finds profitable loans, cuts risks, but lowers acceptance rate.•Explainability tools show different feature uses and key aspects of attractive loans.