Add to Favourites
To login click here
Clara Higuera, lead data scientist at BBVA, discusses the bank’s new machine learning pipeline for early debt recovery and how it allows for earlier and more effective solutions for clients facing financial difficulties. The pipeline includes five models, including one that predicts the probability of a client facing difficulties and another for predicting prolonged default. The use of XG-Boost algorithm has raised concerns about interpretability, but Higuera assures that the models are regularly monitored and can be explained to clients if needed.