| dc.description.abstract | This paper analyzes the different methods of credit risk assessment, considering the techniques of Logistic Regression, widely used for credit risk scoring, and also Machine Learning, which has been explored for various purposes, including in the context of credit scoring. The importance of the theme is evident considering that the economic system depends on credit to promote the expansion of consumption and economic growth. To reach the objective, bibliographic research of a qualitative nature was used in the study. As findings, it is noteworthy that there is no definitive position on which method of credit risk assessment is the best to be used, having different factors for the choice to be made. The fact is that both demonstrate a satisfactory level of accuracy, presenting advantages and disadvantages, depending on the context in which the problem is inserted. Another relevant point, specific to credit in the Brazilian economy, is the level of credit concession that shows expansion over time, maintaining or even reducing the level of default, depending on the period analyzed. This shows that the techniques and processes used by the financial system are following this evolution. As contributions, the paper provides a broad review of the subject, from the role of credit in the economy and its recent evolution in the Brazilian context, to the most used techniques for credit scoring. This research contributes both to the academic environment, by establishing a synthetic consolidation on the topic, and to the professional environment, for professionals in the business of credit risk and related areas. | en |
