Proposal for a credit analysis solution based on the use of artificial intelligence techniques

Abstract

Com certeza. Aqui estão as traduções do texto.

English

With the increase in credit operations, the need to evaluate customer information before making a sale is essential. In this process, the customers' history becomes important for obtaining a faster and more secure response. The use of artificial intelligence techniques to analyze this information and provide the necessary support becomes a possibility. With these techniques, it is possible to train artificial intelligences to assist in decision-making and perform tasks previously carried out by people. This study sought to develop a credit analysis system that uses artificial intelligence techniques to provide a quick and secure decision, minimizing the risks of default. The system simulation was performed using tools such as Python and Sklearn. The GridSearch method was used to find the best model. The simulation results indicated that the use of these techniques increases the chances of identifying customers who are candidates for default. The model's average success rate, analyzing the classification metrics, was 75% for accuracy, 76% for precision, 75% for recall, and 75% for f1-score. The analysis is based on the comparison between the success percentage of customers who were or were not defaulters and were correctly classified by the algorithm. With the inclusion of more detailed information about the customers, the precision will be even higher, making it decisive for the credit analyst's decision-making in more extreme scenarios