Synthetic Mobile Money Transaction Dataset

This dataset comprises synthetic mobile money transaction records, meticulously generated to closely resemble real transaction data for training machine learning models in financial fraud detection. It was produced using the MoMTSim platform, a multi-agent-based simulation specifically tailored to emulate the mobile money financial ecosystem. The simulation inputs are based on real financial data to ensure realism in transaction dynamics. The dataset's fidelity to real-world data has been validated through several statistical methods including the sum of squared errors approach, the Kolmogorov-Smirnov test, and Bland-Altman plots, confirming its accuracy and utility for research and practical applications in detecting fraudulent transactions.