Publicação
Modelos de avaliação para atribuição de microcrédito em Angola
| Resumo: | Various credit scoring models have been used and applied over the years in different countries. These models are used to assess whether a borrower is in a position to be provided with a banking service, usually loans. There are various types of loans, but one that has become more popular in recent years is microcredit. The social function of this type of loan is to help develop the economy and population of less developed countries. By building a credit assessment model, we can use the available data on the borrower to make this assignment, which is the focus of the project. The microcredit mechanism is one of HEFESTO’s many features, a solution that aims to automate, facilitate and reduce the workload of banks in certain services. The pilot project will be tested in Angola and the aim is to adapt it to countries and regions with similar conditions. This project differs from other credit assessment projects in developing countries in that countries in that more emphasis will be placed on demographic, social or psychological variables. These variables reflect both the conditions and the location, rather than just financial variables, since the repayment rate is very high. Data sets A and B are obtained from the same database, but they have different variables; these were 2 data sets obtained using different variable selection methods. These included models such as Multilayer Perceptron, Mixture of Experts and Gradient Boosting. The values obtained by these models were expressed from the F1-Score. |
|---|---|
| Autores principais: | Gonçalves, Bruno Rafael da Luz |
| Assunto: | Avaliação de crédito Microcrédito F1-Score Países em desenvolvimento Teses de mestrado - 2023 |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | português |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | Various credit scoring models have been used and applied over the years in different countries. These models are used to assess whether a borrower is in a position to be provided with a banking service, usually loans. There are various types of loans, but one that has become more popular in recent years is microcredit. The social function of this type of loan is to help develop the economy and population of less developed countries. By building a credit assessment model, we can use the available data on the borrower to make this assignment, which is the focus of the project. The microcredit mechanism is one of HEFESTO’s many features, a solution that aims to automate, facilitate and reduce the workload of banks in certain services. The pilot project will be tested in Angola and the aim is to adapt it to countries and regions with similar conditions. This project differs from other credit assessment projects in developing countries in that countries in that more emphasis will be placed on demographic, social or psychological variables. These variables reflect both the conditions and the location, rather than just financial variables, since the repayment rate is very high. Data sets A and B are obtained from the same database, but they have different variables; these were 2 data sets obtained using different variable selection methods. These included models such as Multilayer Perceptron, Mixture of Experts and Gradient Boosting. The values obtained by these models were expressed from the F1-Score. |
|---|