Authors: Yorgos Goletsis, Themis P. Exarchos, Christos D. Katsis
Addresses: Department of Economics, University of Ioannina, GR45110 Ioannina, Greece. ' Department of Materials Science and Engineering, University of Ioannina, GR45110 Ioannina, Greece. ' Department of Applications of Information Technology in Administration and Economy, Technological Educational Institute of Ionian Islands, GR31100, Lefkada, Greece
Abstract: The application of quantitative techniques for the determination of credit worthiness, i.e., the credit scoring, is a major research field for bankers and academics as it can bring about significant savings to finance institutions whilst minimising their exposure to risk. In the current work, the applicability of recent developments in machine learning techniques is examined; specifically biologically inspired techniques mimicking natural ants, bird flocking and immune system cells are applied. Experimental results are presented on three real world credit scoring datasets. Comparative results with commonly used artificial intelligence and statistical classifiers verify the suitability of the newly examined methods.
Keywords: credit scoring; biologically inspired algorithms; ant colony optimisation; ACO; artificial immune systems; particle swarm optimisation; PSO; classification; neural networks; ANNs; support vector machines; SVM; decision trees; logit; QDA; credit worthiness; machine learning; bio-inspired intelligence.
International Journal of Financial Markets and Derivatives, 2011 Vol.2 No.1/2, pp.32 - 49
Published online: 28 Feb 2015 *Full-text access for editors Access for subscribers Purchase this article Comment on this article