A SOM and GP tool for reducing the dimensionality of a financial distress prediction problem

Authors:
E. Alfaro-Cid;A. M. Mora;J. J. Merelo;A. I. Esparcia-Alcázar;K. Sharman
Affiliations:
Instituto Tecnológico de Informática, Universidad Politécnica de Valencia, Spain;Departamento de Arquitectura y Tecnología de Computadores, Universidad de Granada, Spain;Departamento de Arquitectura y Tecnología de Computadores, Universidad de Granada, Spain;Instituto Tecnológico de Informática, Universidad Politécnica de Valencia, Spain;Instituto Tecnológico de Informática, Universidad Politécnica de Valencia, Spain
Venue:
Evo'08 Proceedings of the 2008 conference on Applications of evolutionary computing
Year:
2008

Citing 8
Cited 0

Self-Organizing Maps

Self-Organizing Maps
Application of Feature Extractive Algorithm to Bankruptcy Prediction

IJCNN '00 Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks (IJCNN'00)-Volume 5 - Volume 5
Biologically Inspired Algorithms for Financial Modelling (Natural Computing Series)

Biologically Inspired Algorithms for Financial Modelling (Natural Computing Series)
Genetic programming for the prediction of insolvency in non-life insurance companies

Computers and Operations Research
The class imbalance problem: A systematic study

Intelligent Data Analysis
Strongly typed genetic programming

Evolutionary Computation
Predicting financial distress: a case study using self-organizing maps

IWANN'07 Proceedings of the 9th international work conference on Artificial neural networks
Bankruptcy analysis with self-organizing maps in learning metrics

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.01

Visualization

Abstract

In order to build prediction models that can be applied to an extensive number of practical cases we need simple models which require the minimum amount of data. The Kohonen's self organizing map (SOM) are usually used to find unknown relationships between a set of variables that describe a problem, and to identify those with higher significance. In this work we have used genetic programming (GP) to produce models that can predict if a company is going to have book losses in the future. In addition, the analysis of the resulting GP trees provides information about the relevance of certain variables when solving the prediction model. This analysis in combination with the conclusions yielded using a SOM have allowed us to reduce significantly the number of variables used to solve the book losses prediction problem while improving the error rates obtained.