Visual analytics para o acompanhamento de modelos de credit scoring

Abstract

Financial institutions use credit Scoring models to predict the default of their customers and assist in decision-making about the granting of credit. As there is a large volume of credit transactions being generated daily and a potential increase in this information with the advent of Open Finance, there is the challenge of being able to monitor this information quickly so we can act in case of loss of performance of these models. Several works found in the literature aim to improve the techniques used in the model construction stage. However, we did not find studies related to monitoring these models. Considering this context, the main objective of this research was to create a Visual Analytics approach to assist in the management of credit models. For this, initially, we carried out a systematic review of the literature on the subject and conducted semi-structured interviews with 13 professionals who have experience in the area. Considering the needs raised with this study, we created a prototype called VACS (Visual Analytics for Tracking Credit Scoring Models). The main contributions of this work are: (a) The results obtained from the systematic review of the literature shows that there is a gap on the subject and allowed us to identify insights into the use of visual analytics and scenario analysis in monitoring those models; (b) The survey of requirements carried out through interviews with specialists, which allowed the recording of how the models are monitored within financial institutions, something that is not disclosed and that can help in the standardization; (c) VACS, which was evaluated by four domain experts who rated it as a very complete and easy-to-use tool; (d) The suggestions collected in the feedback stage, which will contribute to the improvement of VACS and future work.