Extreme Learning Machine to Analyze the Level of Default in Spanish Deposit Institutions

Áreas de investigación:

• Methodology - Machine learning

Año:

2012

Tipo de publicación:

Artículo

Autores:

• Montero-Romero, T.
• López-Martín, M. C.
• Becerra-Alonso, David
• Martínez-Estudillo, Francisco José

Journal:

Revista de Métodos Cuantitativos para la Economía y la Empresa

Volumen:

13

Páginas:

3-23

Nota:

ISSN: 1886-516X

Abstract:

The level of default in financial institutions is a key piece of information in the activity of these organizations and reveals their level of risk. This in turn explains the growing attention given to variables of this kind, during the crisis of these last years. This paper presents a method to estimate the default rate using the non-linear model defined by standard Multilayer Perceptron (MLP) neural networks trained with a novel methodology called Extreme Learning Machine (ELM). The experimental results are promising, and show a good performance when comparing the MLP model trained with the Leverberg-Marquard algorithm.