Detection of credit card fraud with artificial neural networks

In Along with the Internet, digital technologies are frequently used in every moment of our lives. Many transactions that we carry out in monetary terms such as shopping in our daily life are now done digitally. With the developing digitalization in the world, people's lives become easier and people can access different products in a short time. In particular, people can spend and shop quickly and easily without carrying cash in their pockets with a credit card. However, with the increase in the use of credit cards, there are also some security vulnerabilities. Fraudsters can gain unfair advantage by obtaining certain credit card information such as passwords. They can shop with someone else's credit card without permission. These transactions cause substantial financial damage to individuals and institutions. With the increase in the use of credit cards with the developing technology, such credit card fraud is also increasing rapidly. Taking precautions against credit card fraud is a very important issue in order to ensure the safety of people. For this reason, in order to ensure the security of both banks and financial institutions that provide credit card services, it is necessary to prevent credit card fraud and to detect fraud that may occur in credit cards within the scope of combating fraud. In our study, Artificial Neural Networks (ANN) were used to detect credit card fraud transactions. A prediction model that detects fraud in credit card transactions has been developed with ANN. The purpose of this study is to automatically detect abnormal behavior with credit cards. Modeling was done ANN method using the credit card dataset of 284,807 transactions obtained from the Kaggle database. The distribution of 5000 thousand data sets used in modeling was used for 70% training, 15% validation and 15% testing purposes. 3500 samples were used as training data, 750 samples as validation data and the remaining 750 samples as test data. The data preprocessing step was normalized with the MATLAB tool so that the minimum values were 0 and the maximum values were 1. Success and performance rates vary depending on parametric values such as the number of neurons, number of layers, training function and performance function. In this study, it was tried to obtain optimum results by using different layers, different numbers of neurons, and different training and performance functions. In the proposed model used as five layers [28 20 15 10 2], 98.44% success was achieved according to the number of neurons. The training function used is "traincsg" and the performance function is determined as "crossentropy". In a different analysis, 99.14% success was achieved according to the number of neurons in the model using five layers [28 20 15 10 2]. The training function used in this analysis was "trainbr" and the performance function was chosen as "mse (mean squalre error)".