Counterfeit Diagnosis Using Artificial Neural Network

Abstract

Abstract: One of the best payment methods that clients may use to conduct transactions effortlessly and boost their purchasing power is a credit card. In addition, it is incredibly simple to use and provides a number of benefits, like cashback and reward points. However, the primary issue is credit card fraud, which is fast rising every day. The most common type of identity theft worldwide is still credit card fraud. The use of credit cards has grown too dangerous in recent years. In reality, there are many other kinds of credit card fraud, including phishing and vishing, keystroke logging, POS fraud, application fraud, loss of card or theft, and POS fraud and vishing. However, the person should take safety steps and safeguards to protect his or her money. The vast bulk of this fraud is being committed by organized crime rings, whose operations have been industrialized and computerized. So how may these frauds be found? Utilizing machine learning algorithms is the solution. Conventional fraud detection is another method, however, machine learning algorithms are much more accurate and exact. The basic objective is to create a model that can foretell whether a Transaction will be fraudulent or not. In this project, predictive models utilized in this research include Artificial Neural Networks, Random Forests, Support Vector Machines, KNN, Decision Trees, Gaussian Naive Bayes, and Logistic Regression. The results from each of these models are evaluated for accuracy, and the best model is selected.