Machine learning aplicado no problema de perdas com créditos de uma distribuidora de energia elétrica

Abstract

Allowance for Doubtful Accounts (AFDA) in companies is an attractive field for investigation due to the percentage of profits it represents. The objective of this work is to find a machine learning model to predict which day the customer will pay the invoice in order to maximize the company’s profit. To evaluate the proposed methodology, experiments were carried out using real data from customer invoices. The results of the models were compared with each other and a statistical analysis was carried out to verify if there was a significant difference between them. The results indicate that it is promising to apply the proposed model to the problem of Estimated Losses on Loan Losses.