Aplicación de Modelos de Machine Learning para la Predicción de Quiebra en el Mercado Financiero Peruano Durante el Periodo 2014-2024

The early prediction of corporate bankruptcy is a key challenge for economic stability and risk management in emerging markets. This study explores the application of various Machine Learning techniques to anticipate the probability of insolvency in companies operating in the Peruvian market. The ma...

Descripción completa

Detalles Bibliográficos
Autor: Leon Campos, Piero Martin
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:Perú
Institución:Universidad Nacional Mayor de San Marcos
Repositorio:Revistas - Universidad Nacional Mayor de San Marcos
Idioma:español
OAI Identifier:oai:revistasinvestigacion.unmsm.edu.pe:article/31334
Acceso en línea:https://revistasinvestigacion.unmsm.edu.pe/index.php/rpcsis/article/view/31334
Access Level:acceso abierto
Palabra clave:predicción de quiebras
Machine Learning (ML)
riesgo financiero
support vector machine
redes neuronales
bankruptcy prediction
machine learning
financial risk
neural networks
Descripción
Sumario:The early prediction of corporate bankruptcy is a key challenge for economic stability and risk management in emerging markets. This study explores the application of various Machine Learning techniques to anticipate the probability of insolvency in companies operating in the Peruvian market. The main objective is to compare the predictive capacity of both classical and modern models, including multiple discriminant analysis, logistic regression, Support Vector Machines, Random Forest, Bagging, Boosting, and neural networks, using accounting and financial information. The methodology involved building a database of financial series from Peruvian firms, normalizing ratios, splitting the data into training and testing sets, and evaluating performance through widely used classification metrics such as accuracy, recall, F1 score, Type I and II errors, and AUC. The results show that ensemble-based models, particularly Boosting, Support Vector Machines, and Random Forest, outperform traditional techniques, highlighting the usefulness of machine learning approaches in strengthening financial risk management in local contexts. These findings provide practical implications for both financial institutions and regulators.