Tracking Fraudulent Transactions in Credit Cards using Logistic Regression

Abstract:- E-commerce has reshaped the global trade front, now becoming an all-important tool for organizations, businesses, and governments to increase effectiveness. A primary reason driving the growth of e-commerce is the ease of card transactions through the Internet. However, with the spread of digital payments, cybercrime cases and cybersecurity issues also increased. Fraud financing deliberately denies the victim’s rights or gains unlawful monetary advantages. Thus, detecting fraud is one of the primary concerns for banks and other financial institutions. ML is emerging as an innovative solution in detecting credit card fraud, surpassing conventional techniques by identifying complex patterns within big data. By analyzing user behavior, payment methods, and transaction patterns, it can predict and prevent abnormal activities. This study assesses the capability of ML algorithms to identify fraudulent credit card transactions through the application of algorithms like Logistic Regression (LR), Random Forest (RF), and K-Nearest Neighbors (KNN). Through credit card transactions in Kaggle data, accuracy, precision, and F1-score assessments are evaluated on these models. According to the outcome, the model that depicted maximum accuracy is LR while attaining 94% for fraud detection accuracy. This shows the potential of ML techniques to enhance fraud detection systems and provide more security and efficiency in digital payment systems. The findings also highlight the importance of embracing advanced analytics in combating financial fraud and securing the ever-changing ecommerce landscape.