ADAPTIVE SYSTEM OF THE CREDITWORTHINESS EVALUATION CONSTRUCTED ON THE BASIS OF ARTIFICIAL NEURAL NETWORKS

Dorota Witkowska
Additional contact information
Dorota Witkowska: Technical University of Lodz

No 230, Computing in Economics and Finance 2000 from Society for Computational Economics

Abstract: Artificial neural networks are nonlinear models whose parameters (weights) are estimated during so called training procedure. Applying the back propagation algorithm (which is the most popular method of supervised learning), the computed output (for the initial, usually randomly chosen weights) is compared to the known output. If the generated output is correct, then nothing more is necessary. If the computed output is incorrect, then the weights are adjusted so as to make the computed output closer to the known output. Training procedure runs in a certain (usually determined by the ANN constructor) number of iterations. In many applications we deal with the objective (cost - criterion) function with many local and global minima. From the practical point of view we are interested in the localization of all global minima. The global optimization methods that have been developing rapidly in the last years include genetic algorithms which appeared to be universal, flexible and efficient. Genetic algorithms (GA) are modern successors of Monte Carlo search methods, and they belong to the class of stochastic optimization algorithms. GA are usually a compromise between searching the whole set of feasible solutions and local optimization. But with high probability they lead to the global minimum. The aim of this paper is to construct a classification system, based on neural networks, of clients of the financial institutions. This system is to recognize clients automatically thanks to the experience stored in the training set.

Date: 2000-07-05
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:sce:scecf0:230

Access Statistics for this paper

More papers in Computing in Economics and Finance 2000 from Society for Computational Economics CEF 2000, Departament d'Economia i Empresa, Universitat Pompeu Fabra, Ramon Trias Fargas, 25,27, 08005, Barcelona, Spain. Contact information at EDIRC.
Bibliographic data for series maintained by Christopher F. Baum ().