“The use of Machine Learning (ML) models is gaining traction in finance due to their better predictive capacity compared to traditional statistical techniques…One of the use cases with greater potential is its application to credit underwriting and scoring, since by having better predictive capacity, ML models allow better estimates of the probability of default and therefore could result in more accurate credit scores. But this improvement in predictive performance does not come without risk.”

This white paper creates a framework for using synthetic data sets to assess the accuracy of interpretability techniques as applied to machine learning models in finance. The authors controlled actual feature importance using a synthetic data set and then compared the outputs of two popular interpretability techniques to determine which was better at identifying relevant features, finding variation in results.