PSO-based Linear SVM Classifier Selection for Credit Risk Evaluation Modeling Process

Paulius Danenas, Gintautas Garsva

2012

Abstract

A research on credit risk evaluation modelling using linear Support Vector Machines (SVM) classifiers is proposed in this paper. The classifier selection is automated using Particle Swarm Optimization technique. Sliding window approach is applied for testing classifier performance, together with other techniques such as discriminant analysis based scoring for evaluation of financial instances and correlation-based feature selection. The developed classifier is applied and tested on real bankruptcy data showing promising results.

Paper Citation

in Harvard Style

Danenas P. and Garsva G. (2012). PSO-based Linear SVM Classifier Selection for Credit Risk Evaluation Modeling Process . In Proceedings of the 14th International Conference on Enterprise Information Systems - Volume 1: ICEIS, ISBN 978-989-8565-10-5, pages 338-341. DOI: 10.5220/0004006403380341

in Bibtex Style

@conference{iceis12,
author={Paulius Danenas and Gintautas Garsva},
title={PSO-based Linear SVM Classifier Selection for Credit Risk Evaluation Modeling Process},
booktitle={Proceedings of the 14th International Conference on Enterprise Information Systems - Volume 1: ICEIS,},
year={2012},
pages={338-341},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0004006403380341},
isbn={978-989-8565-10-5},
}

in EndNote Style

TY - CONF
JO - Proceedings of the 14th International Conference on Enterprise Information Systems - Volume 1: ICEIS,
TI - PSO-based Linear SVM Classifier Selection for Credit Risk Evaluation Modeling Process
SN - 978-989-8565-10-5
AU - Danenas P.
AU - Garsva G.
PY - 2012
SP - 338
EP - 341
DO - 10.5220/0004006403380341