To grant a bank credit or not is very important for effective bank management. A standard approach of computer decision support systems is based on spread sheets. The artificial neural network approach gives more possibilities to analyze credit decisions, to classify entities applying for credits according to their credibility, including different groups of risk. In this work, the authoress presents a review of various neural network topologies appliance and net trained using various algorithms for dichotomous and polytomic classification. Classification errors were compared and the most effective net was determined. Advantages and disadvantages of described method were shown, which indicate the right appliance of artificial neural networks for the analysis of loan debtors. The usage of artificial networks can rationalize and speed up the process of granting credits, as well as provide a basis for a secondary verification of refused applications.
Financed by the National Centre for Research and Development under grant No. SP/I/1/77065/10 by the strategic scientific research and experimental development program:
SYNAT - “Interdisciplinary System for Interactive Scientific and Scientific-Technical Information”.