Fraud_Detection_ML: Machine Learning Based on Online Payment Fraud Detection

Abstract

Available online Online payment fraud detection is crucial for safeguarding e-commerce transactions against sophisticated fraudsters who exploit system vulnerabilities. This paper proposes an efficient framework for predicting online payment fraud, employing six diverse machine learning algorithms, namely constant, CN7Rule induction, KNN, Tree, Random Forest, Gradient boosting, SVM, Logistic regression, Naive Bayes, Ada boost, Neural network, and stochastic gradient descent, on three distinct datasets. The gradient-boosting algorithm consistently outperformed others through rigorous testing, achieving an impressive accuracy rate of 99.7%. This algorithm demonstrated resilience across various testing scenarios, establishing itself as the most effective online payment fraud detection solution. With the highest accuracy score of 99.7% in all testing phases, gradient boosting is optimal for preemptive measures against fraudulent activities in electronic transactions, providing a robust defense mechanism for e-commerce platforms. 1. Introduction Online payment fraud detection is a process that prevents fraudulent activities in online transactions. It involves device fingerprinting, geolocation, behavioral analysis, transaction tracking, and two-factor authentication. Machine learning and AI algorithms continuously adapt to new fraud strategies, and cooperation between payment service providers and financial institutions is beneficial [1]. Online payment fraud is a problem that arises from dishonest and illegal actions taken during electronic transactions. Unauthorized transactions, identity theft, compromised payment credentials, phishing, social engineering, insufficient security protocols, account takeover, difficulties with cross-border transactions, and risks associated with developing technologies are some of the major problems [2]. Machine learning is an evolving branch of computational algorithms designed to emulate human intelligence by learning from the surrounding environment. They are considered the working horse in the new era of big data. Techniques based on machine learning have been applied successfully in diverse fields ranging from pattern recognition, computer vision, spacecraft engineering, finance, entertainment, and computational biology to biomedical and medical applications [3]. Machine learning plays a crucial role in addressing the challenge of online payment fraud by enabling automated, data-driven fraud detection and prevention systems. Here's how machine learning is applied to combat online payment fraud [4]. The main contribution of this paper follows: we use six algorithms, and we made predictions for online payment fraud; we use cross-validation (10), training (80%), and testing (20), and the best algorithm was gradient boosting with Accuracy (0.997). The rest of the paper can be organized as follows: Machine learning is an effective technique when it comes to identifying and stopping online payment fraud. It can analyze vast data volumes, spot trends, and