Application of Ann in Australian Credit Card Approval

Nowadays, credit cards became popular in both developed and developing
countries. For lenders, more credit card applications mean more complex decision making process.
The purpose of this research was to examine how accurate and reliable results Neural Networks
(ANN) can provide for Credit Card approval. Research was done using Australian Credit Card
Approval Dataset which has interesting mixture of data and 690 instances. Thirty experiments
were done using two and three hidden layers, and back-propagation training method. Results
indicated high level of reliability and accuracy of 87 % which is good empirical evidence that ANN
can be used as reliable tool for credit card approval