Skip to Main Content
In this study we undertook the credit card fraud detection problem of a bank and tried to improve the performance of an existing solution. In doing so, we did not undertake the typical objective of maximizing the number of correctly classified transactions but rather we defined a new objective function where the misclassification costs are variable and thus, correct classification of some transactions are more important than correctly classifying the others. For this purpose we made an application of genetic algorithms which is a novel one in the related literature both in terms of the application domain and the cross-over operator used. The algorithm is applied to real life data where the savings obtained are almost three times the current practice.