The aim is to explore the contribution that artificial intelligence can make in predicting consumer over-indebtedness using Machine Learning (ML) to develop descriptive and predictive models to understand the factors that influence over-indebtedness of Portuguese consumers. The project goals is to (1) characterize and describe the over-indebtedness of Portuguese consumers using unsupervised ML; (2) create reliable models of supervised ML to help predict the factors that influence over-indebtedness; (3) develop interventions to help in the Alternative Dispute Resolution of consumer debt.

Based on the preliminary results NOVA IMS minimizes is able to predict 9 cases of over-indebtment over 10, with antecedence!

Impact:

1. Provide better counseling for indebted consumers;
2. Anticipate the risk of future cases of over-indebtedness;
3. Improve data, with the help of Directorate General for Consumer Affairs.