Advantages of the PaySim Simulator for Improving Financial Fraud Controls

Abstract

Financial transactions lay up the foundation of modern society. Unfortunately, illicit abuse of the financial system is pervasive. Fraud controls aim to detect these suspicious activities, but they require deep analysis to model their efficacy and value proposition. Due to the private nature and scale of these financial transactions, this analysis is often performed in hindsight. Financial institutions lack of information, due to the hidden fraud problem, to properly set and tune their fraud controls systems. This is probably one of the reasons we are losing the war against crime. This paper presents PaySim, a cutting edge agent-based model that simulates financial fraud scenarios to improve current fraud controls. PaySim uses aggregated anonymized data from a real financial dataset to generate synthetic data that closely resembles the transactions dynamics, statistical properties and causal dynamics observed in the original dataset, while incorporating any malicious behaviour of interest. Using an agent-based framework specifically designed to cover the demands of financial simulation and the application of mathematical statistics, we leverage a real-life scenario based on a known fraud scheme to demonstrate the advantage of simulated data over real-world data when setting adequate controls for fraud detection.

Predicting Mobile Money Transaction Fraud using Machine Learning Algorithms

“…In many cases, mobile money services are being used to facilitate money laundering by allowing criminals to quickly and easily move large amounts of cash without raising suspicion. For example, recent studies found that a significant proportion of mobile money users have been involved in a transaction that could be considered suspicious 8,14,44,45 . Considering these findings, more must be done to prevent mobile money from being used for illegal purposes.…”

“…The data came from Lopez-Rojas and his colleagues' work, who use MABS to develop agents representing clients and merchants in PaySim and customers and salesmen in RetSim 43,44 . The data is simulated and uses a realworld scenario based on a well-known fraud scheme to demonstrate the superiority of simulated data over real-world data when establishing adequate controls for fraud detection 45 . The usual behaviour was derived from the behaviour that was observed in the data collected.…”

Unmasking Banking Fraud: Unleashing the Power of Machine Learning and Explainable AI (XAI) on Imbalanced Data

“…From those papers [11,12,16], Logistic Regression showed better results than the other methods. In [17,18], a simulator named PaySim was used to evaluate several methods to prevent fraud based on thresholds [19] on a mobile money service. The authors of [20] employed a decentralized data mining methodology to tackle the issues of imbalanced credit card distributions.…”

Predicting mobile money transaction fraud using machine learning algorithms

“…28,37 The data is simulated and utilizes a real-world scenario based on a well-known fraud scheme to highlight the superiority of simulated data over real-world data when establishing adequate controls for fraud detection. 29 The agents' rules governing transactions and interactions between consumers and salespeople or customers and merchants are derived from observed behavioral patterns in the collected data. Some agents were specifically designed to engage in fraudulent activities based on actual fraud patterns.…”