Efficient Credit Card Fraud Detection Based on Binary Logistic Regression

Abstract

With the rapid increase in credit card usage, instances of credit card fraud are also on the rise. The aim of this paper is to design a credit card fraud detection model using binary logistic regression. By using effective detection techniques, the model increases detection accuracy, safeguarding consumer interests and preserving financial market stability. The findings demonstrate that the binary logistic regression model developed for this investigation has a 93.9% accuracy rate in identifying credit card fraud. Important metrics like recall rate and accuracy rate performed exceptionally well, reaching 93.1% and 94.5%, respectively. The model significantly lowers false positives and incorrect assessments in addition to being very good at spotting fraudulent transactions. In addition to offering a reference for resolving other financial fraud detection issues, the paper presents a new method of credit card fraud detection. By improving the model and incorporating additional data characteristics, its performance and applicability can be further enhanced to provide financial institutions with stronger support against future fraud threats.