Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Development of New Methods for Detection and Control of Credit Card Fraud Attacks

Alexander Stotsky

TL;DR

This paper tackles the problem of detecting and controlling credit card fraud attacks in real time, where attacks manifest as rapid sequences of fraudulent transactions. It combines a moving-window confidence interval with exponential forgetting to adapt to evolving cardholder behavior and detect outliers, with a fraud risk scoring approach to identify the onset of an attack based on multiple risk factors. Detection is followed by control strategies that adjust the interval and impose mitigations (e.g., block/limit high-risk transactions, constrain MCCs, and require extra verification) to minimize losses. The approach is evaluated conceptually on a dataset of approximately $24.4$ million transactions from IBM, and the authors call for robust real-time algorithms and simulation tools to prepare for emerging attack types in crisis scenarios.

Abstract

Credit card fraud causes significant financial losses and frequently occurs as fraud attack, defined as short-term sequence of fraudulent transactions associated with high transaction rates and amounts, business areas historically tied to fraud, unusual transaction times and locations and different types of errors. Confidence interval method in the moving window with exponential forgetting is proposed in this report which allows to capture recent changes in the shopping behaviour of the cardholder, detect fraudulent amounts and mitigate the attack. Fraud risk scoring method is used for estimation of the intensity of the fraudulent activity via monitoring of the transaction rates, merchant category codes, times and some other factors for detection of the start of the attack. The development and verification are based on detailed analysis of the transaction patterns from the dataset, which represents an extensive collection of around 24.4 million credit card transactions from IBM financial database. Recommendations for further development of the detection techniques are also presented.

Development of New Methods for Detection and Control of Credit Card Fraud Attacks

TL;DR

This paper tackles the problem of detecting and controlling credit card fraud attacks in real time, where attacks manifest as rapid sequences of fraudulent transactions. It combines a moving-window confidence interval with exponential forgetting to adapt to evolving cardholder behavior and detect outliers, with a fraud risk scoring approach to identify the onset of an attack based on multiple risk factors. Detection is followed by control strategies that adjust the interval and impose mitigations (e.g., block/limit high-risk transactions, constrain MCCs, and require extra verification) to minimize losses. The approach is evaluated conceptually on a dataset of approximately million transactions from IBM, and the authors call for robust real-time algorithms and simulation tools to prepare for emerging attack types in crisis scenarios.

Abstract

Credit card fraud causes significant financial losses and frequently occurs as fraud attack, defined as short-term sequence of fraudulent transactions associated with high transaction rates and amounts, business areas historically tied to fraud, unusual transaction times and locations and different types of errors. Confidence interval method in the moving window with exponential forgetting is proposed in this report which allows to capture recent changes in the shopping behaviour of the cardholder, detect fraudulent amounts and mitigate the attack. Fraud risk scoring method is used for estimation of the intensity of the fraudulent activity via monitoring of the transaction rates, merchant category codes, times and some other factors for detection of the start of the attack. The development and verification are based on detailed analysis of the transaction patterns from the dataset, which represents an extensive collection of around 24.4 million credit card transactions from IBM financial database. Recommendations for further development of the detection techniques are also presented.

Paper Structure

This paper contains 5 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Confidence Interval for Detection & Control
  3. Detection of the Start of Attack: Fraud Risk Scoring Method
  4. Control of Fraud Attack in the Case of Uncertainty
  5. Conclusion

Figures (1)

  • Figure 1: Detection and control of the fraud attack is presented in this Figure. The confidence is plotted with green lines. The transaction amount is plotted with the blue line, weighted moving average is plotted with the black line and fraudulent amounts are plotted with red round signs. Fraudulent transaction is detected if shopping amount exceeds the threshold. The first two transactions in the attack were flagged as fraudulent due small inter-transaction time gap, technical glitch and unusual time. The confidence interval was assigned to zero in each step of moving window and several online fraudulent transactions with high rates and amounts, which took place at unusual time with technical glitches and CVV errors have been blocked and the fraud attack was prevented. Similar procedure was applied to the second fraud attack shown in the Figure. The ability of the detection system to recover after the attacks is simply associated with resetting of the length of the confidence interval.