Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina
Ano de defesa: | 2018 |
---|---|
Autor(a) principal: | |
Orientador(a): | |
Banca de defesa: | |
Tipo de documento: | Dissertação |
Tipo de acesso: | Acesso aberto |
Idioma: | por |
Instituição de defesa: |
Universidade Católica de Brasília
|
Programa de Pós-Graduação: |
Programa Stricto Sensu em Economia de Empresas
|
Departamento: |
Escola de Gestão e Negócios
|
País: |
Brasil
|
Palavras-chave em Português: | |
Área do conhecimento CNPq: | |
Resumo em Inglês: | The credit risk assessment has a relevant role for financial institutions because it is associated with possible losses and has a large impact on the balance sheets. Although there are several researches on applications of machine learning and finance models, a study is still lacking that integrates available knowledge about credit risk assessment. This paper aims at specifying the machine learning model of the probability of default of publicly traded companies present in the Bovespa Index (corporations) and, based on the estimations of the model, to obtain risk assessment metrics based on risk letters. We converged methodologies verified in the literature and we estimated models that comprise fundamentalist (balance sheet) and governance data, macroeconomic and even variables resulting from the application of the proprietary model of KMV credit risk assessment. We test the XGboost and LinearSVM algorithms, which have very different characteristics among them, but are potentially useful to the problem. Parameter Grids were performed to identify the most representative variables and to specify the best performing model. The model selected was XGboost, and performance was very similar to the results obtained for the North American stock market in analogous research. The estimated credit ratings suggest that they are more sensitive to the economic and financial situation of the companies than that verified by traditional Rating Agencies. |
Link de acesso: | https://bdtd.ucb.br:8443/jspui/handle/tede/2432 |
Resumo: | The credit risk assessment has a relevant role for financial institutions because it is associated with possible losses and has a large impact on the balance sheets. Although there are several researches on applications of machine learning and finance models, a study is still lacking that integrates available knowledge about credit risk assessment. This paper aims at specifying the machine learning model of the probability of default of publicly traded companies present in the Bovespa Index (corporations) and, based on the estimations of the model, to obtain risk assessment metrics based on risk letters. We converged methodologies verified in the literature and we estimated models that comprise fundamentalist (balance sheet) and governance data, macroeconomic and even variables resulting from the application of the proprietary model of KMV credit risk assessment. We test the XGboost and LinearSVM algorithms, which have very different characteristics among them, but are potentially useful to the problem. Parameter Grids were performed to identify the most representative variables and to specify the best performing model. The model selected was XGboost, and performance was very similar to the results obtained for the North American stock market in analogous research. The estimated credit ratings suggest that they are more sensitive to the economic and financial situation of the companies than that verified by traditional Rating Agencies. |
id |
UCB_9df3c72aeecdb2c22690746989d8ba68 |
---|---|
oai_identifier_str |
oai:bdtd.ucb.br:tede/2432 |
network_acronym_str |
UCB |
network_name_str |
Biblioteca Digital de Teses e Dissertações da UCB |
spelling |
Silva Filho, Osvaldo Cândido dahttp://lattes.cnpq.br/3691103797905606http://lattes.cnpq.br/7136224867022033Gregório, Rafael Leite2018-08-08T13:33:24Z2018-07-09GREGÓRIO, Rafael Leite. Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina. 2018. 70 f. Dissertação (Programa Stricto Sensu em Economia de Empresas) - Universidade Católica de Brasília, Brasília, 2018.https://bdtd.ucb.br:8443/jspui/handle/tede/2432The credit risk assessment has a relevant role for financial institutions because it is associated with possible losses and has a large impact on the balance sheets. Although there are several researches on applications of machine learning and finance models, a study is still lacking that integrates available knowledge about credit risk assessment. This paper aims at specifying the machine learning model of the probability of default of publicly traded companies present in the Bovespa Index (corporations) and, based on the estimations of the model, to obtain risk assessment metrics based on risk letters. We converged methodologies verified in the literature and we estimated models that comprise fundamentalist (balance sheet) and governance data, macroeconomic and even variables resulting from the application of the proprietary model of KMV credit risk assessment. We test the XGboost and LinearSVM algorithms, which have very different characteristics among them, but are potentially useful to the problem. Parameter Grids were performed to identify the most representative variables and to specify the best performing model. The model selected was XGboost, and performance was very similar to the results obtained for the North American stock market in analogous research. The estimated credit ratings suggest that they are more sensitive to the economic and financial situation of the companies than that verified by traditional Rating Agencies.A avaliação do risco de crédito tem papel relevante para as instituições financeiras por estar associada a possíveis perdas que podem gerar grande impacto nos balanços. Embora existam várias pesquisas sobre aplicações de modelos de aprendizado de máquina e finanças, ainda não há estudo que integre o conhecimento disponível sobre avaliação de risco de crédito. Este trabalho visa especificar modelo de aprendizado de máquina da probabilidade de descumprimento de empresas de capital aberto presentes no Índice Bovespa (corporações) e, fruto das estimações do modelo, obter métrica de avaliação de risco baseada em letras (ratings) de risco. Convergiu-se metodologias verificadas na literatura e estimou-se modelos que compreendem componentes fundamentalistas (de balanço) e de governança corporativa, macroeconômicos e ainda variáveis produto da aplicação do modelo proprietário de avaliação de risco de crédito KMV. Testou-se os algoritmos XGboost e LinearSVM, os quais possuem características bastante distintas entre si, mas são potencialmente úteis ao problema exposto. Foram realizados Grids de parâmetros para identificação das variáveis mais representativas e para a especificação do modelo com melhor desempenho. O modelo selecionado foi o XGboost, tendo sido observado desempenho bastante semelhante aos resultados obtidos para o mercado de ações norte-americano em pesquisa análoga. Os ratings de crédito estimados mostram-se mais sensíveis à situação econômico-financeira das empresas ante o verificado por agências de rating tradicionais.Submitted by Sara Ribeiro (sara.ribeiro@ucb.br) on 2018-08-08T13:33:03Z No. of bitstreams: 1 RafaelLeiteGregorioDissertacao2018.pdf: 1382550 bytes, checksum: 9c6e4f1d3c561482546aca581262b92b (MD5)Approved for entry into archive by Sara Ribeiro (sara.ribeiro@ucb.br) on 2018-08-08T13:33:24Z (GMT) No. of bitstreams: 1 RafaelLeiteGregorioDissertacao2018.pdf: 1382550 bytes, checksum: 9c6e4f1d3c561482546aca581262b92b (MD5)Made available in DSpace on 2018-08-08T13:33:24Z (GMT). No. of bitstreams: 1 RafaelLeiteGregorioDissertacao2018.pdf: 1382550 bytes, checksum: 9c6e4f1d3c561482546aca581262b92b (MD5) Previous issue date: 2018-07-09application/pdfhttps://bdtd.ucb.br:8443/jspui/retrieve/5831/RafaelLeiteGregorioDissertacao2018.pdf.jpgporUniversidade Católica de BrasíliaPrograma Stricto Sensu em Economia de EmpresasUCBBrasilEscola de Gestão e NegóciosSVMXGboostRisco de créditoRatings de créditoDefault probabilityCredit riskCNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIAModelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquinainfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisinfo:eu-repo/semantics/openAccessreponame:Biblioteca Digital de Teses e Dissertações da UCBinstname:Universidade Católica de Brasíliainstacron:UCBTHUMBNAILRafaelLeiteGregorioDissertacao2018.pdf.jpgRafaelLeiteGregorioDissertacao2018.pdf.jpgimage/jpeg4767https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/4/RafaelLeiteGregorioDissertacao2018.pdf.jpgc23cb3d79526b17d57c0be2f224f9dceMD54TEXTRafaelLeiteGregorioDissertacao2018.pdf.txtRafaelLeiteGregorioDissertacao2018.pdf.txttext/plain123884https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/3/RafaelLeiteGregorioDissertacao2018.pdf.txt1674e5e51fc22a5d533014aec794f12eMD53ORIGINALRafaelLeiteGregorioDissertacao2018.pdfRafaelLeiteGregorioDissertacao2018.pdfapplication/pdf1382550https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/2/RafaelLeiteGregorioDissertacao2018.pdf9c6e4f1d3c561482546aca581262b92bMD52LICENSElicense.txtlicense.txttext/plain; charset=utf-81905https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/1/license.txt75558dcf859532757239878b42f1c2c7MD51tede/2432oai:bdtd.ucb.br:tede/24322018-08-09 01:10:38.819Biblioteca Digital de Dissertações da Universidade Católica de Brasília - UCBsdi@ucb.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 |
dc.title.por.fl_str_mv |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
title |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
spellingShingle |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina Gregório, Rafael Leite SVM XGboost Risco de crédito Ratings de crédito Default probability Credit risk CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA |
title_short |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
title_full |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
title_fullStr |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
title_full_unstemmed |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
title_sort |
Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina |
author |
Gregório, Rafael Leite |
author_facet |
Gregório, Rafael Leite |
author_role |
author |
dc.contributor.advisor1.fl_str_mv |
Silva Filho, Osvaldo Cândido da |
dc.contributor.advisor1Lattes.fl_str_mv |
http://lattes.cnpq.br/3691103797905606 |
dc.contributor.authorLattes.fl_str_mv |
http://lattes.cnpq.br/7136224867022033 |
dc.contributor.author.fl_str_mv |
Gregório, Rafael Leite |
contributor_str_mv |
Silva Filho, Osvaldo Cândido da |
dc.subject.por.fl_str_mv |
SVM XGboost Risco de crédito Ratings de crédito Default probability Credit risk |
topic |
SVM XGboost Risco de crédito Ratings de crédito Default probability Credit risk CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA |
dc.subject.cnpq.fl_str_mv |
CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA |
dc.description.abstract.eng.fl_txt_mv |
The credit risk assessment has a relevant role for financial institutions because it is associated with possible losses and has a large impact on the balance sheets. Although there are several researches on applications of machine learning and finance models, a study is still lacking that integrates available knowledge about credit risk assessment. This paper aims at specifying the machine learning model of the probability of default of publicly traded companies present in the Bovespa Index (corporations) and, based on the estimations of the model, to obtain risk assessment metrics based on risk letters. We converged methodologies verified in the literature and we estimated models that comprise fundamentalist (balance sheet) and governance data, macroeconomic and even variables resulting from the application of the proprietary model of KMV credit risk assessment. We test the XGboost and LinearSVM algorithms, which have very different characteristics among them, but are potentially useful to the problem. Parameter Grids were performed to identify the most representative variables and to specify the best performing model. The model selected was XGboost, and performance was very similar to the results obtained for the North American stock market in analogous research. The estimated credit ratings suggest that they are more sensitive to the economic and financial situation of the companies than that verified by traditional Rating Agencies. |
dc.description.abstract.por.fl_txt_mv |
A avaliação do risco de crédito tem papel relevante para as instituições financeiras por estar associada a possíveis perdas que podem gerar grande impacto nos balanços. Embora existam várias pesquisas sobre aplicações de modelos de aprendizado de máquina e finanças, ainda não há estudo que integre o conhecimento disponível sobre avaliação de risco de crédito. Este trabalho visa especificar modelo de aprendizado de máquina da probabilidade de descumprimento de empresas de capital aberto presentes no Índice Bovespa (corporações) e, fruto das estimações do modelo, obter métrica de avaliação de risco baseada em letras (ratings) de risco. Convergiu-se metodologias verificadas na literatura e estimou-se modelos que compreendem componentes fundamentalistas (de balanço) e de governança corporativa, macroeconômicos e ainda variáveis produto da aplicação do modelo proprietário de avaliação de risco de crédito KMV. Testou-se os algoritmos XGboost e LinearSVM, os quais possuem características bastante distintas entre si, mas são potencialmente úteis ao problema exposto. Foram realizados Grids de parâmetros para identificação das variáveis mais representativas e para a especificação do modelo com melhor desempenho. O modelo selecionado foi o XGboost, tendo sido observado desempenho bastante semelhante aos resultados obtidos para o mercado de ações norte-americano em pesquisa análoga. Os ratings de crédito estimados mostram-se mais sensíveis à situação econômico-financeira das empresas ante o verificado por agências de rating tradicionais. |
description |
The credit risk assessment has a relevant role for financial institutions because it is associated with possible losses and has a large impact on the balance sheets. Although there are several researches on applications of machine learning and finance models, a study is still lacking that integrates available knowledge about credit risk assessment. This paper aims at specifying the machine learning model of the probability of default of publicly traded companies present in the Bovespa Index (corporations) and, based on the estimations of the model, to obtain risk assessment metrics based on risk letters. We converged methodologies verified in the literature and we estimated models that comprise fundamentalist (balance sheet) and governance data, macroeconomic and even variables resulting from the application of the proprietary model of KMV credit risk assessment. We test the XGboost and LinearSVM algorithms, which have very different characteristics among them, but are potentially useful to the problem. Parameter Grids were performed to identify the most representative variables and to specify the best performing model. The model selected was XGboost, and performance was very similar to the results obtained for the North American stock market in analogous research. The estimated credit ratings suggest that they are more sensitive to the economic and financial situation of the companies than that verified by traditional Rating Agencies. |
publishDate |
2018 |
dc.date.accessioned.fl_str_mv |
2018-08-08T13:33:24Z |
dc.date.issued.fl_str_mv |
2018-07-09 |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
status_str |
publishedVersion |
format |
masterThesis |
dc.identifier.citation.fl_str_mv |
GREGÓRIO, Rafael Leite. Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina. 2018. 70 f. Dissertação (Programa Stricto Sensu em Economia de Empresas) - Universidade Católica de Brasília, Brasília, 2018. |
dc.identifier.uri.fl_str_mv |
https://bdtd.ucb.br:8443/jspui/handle/tede/2432 |
identifier_str_mv |
GREGÓRIO, Rafael Leite. Modelo híbrido de avaliação de risco de crédito para corporações brasileiras com base em algoritmos de aprendizado de máquina. 2018. 70 f. Dissertação (Programa Stricto Sensu em Economia de Empresas) - Universidade Católica de Brasília, Brasília, 2018. |
url |
https://bdtd.ucb.br:8443/jspui/handle/tede/2432 |
dc.language.iso.fl_str_mv |
por |
language |
por |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
Universidade Católica de Brasília |
dc.publisher.program.fl_str_mv |
Programa Stricto Sensu em Economia de Empresas |
dc.publisher.initials.fl_str_mv |
UCB |
dc.publisher.country.fl_str_mv |
Brasil |
dc.publisher.department.fl_str_mv |
Escola de Gestão e Negócios |
publisher.none.fl_str_mv |
Universidade Católica de Brasília |
dc.source.none.fl_str_mv |
reponame:Biblioteca Digital de Teses e Dissertações da UCB instname:Universidade Católica de Brasília instacron:UCB |
reponame_str |
Biblioteca Digital de Teses e Dissertações da UCB |
collection |
Biblioteca Digital de Teses e Dissertações da UCB |
instname_str |
Universidade Católica de Brasília |
instacron_str |
UCB |
institution |
UCB |
bitstream.url.fl_str_mv |
https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/4/RafaelLeiteGregorioDissertacao2018.pdf.jpg https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/3/RafaelLeiteGregorioDissertacao2018.pdf.txt https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/2/RafaelLeiteGregorioDissertacao2018.pdf https://bdtd.ucb.br:8443/jspui/bitstream/tede/2432/1/license.txt |
bitstream.checksum.fl_str_mv |
c23cb3d79526b17d57c0be2f224f9dce 1674e5e51fc22a5d533014aec794f12e 9c6e4f1d3c561482546aca581262b92b 75558dcf859532757239878b42f1c2c7 |
bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 MD5 MD5 |
repository.name.fl_str_mv |
Biblioteca Digital de Dissertações da Universidade Católica de Brasília - UCB |
repository.mail.fl_str_mv |
sdi@ucb.br |
_version_ |
1643299889520574464 |