金融詐欺預測分析

Financial Statement Fraud Detection

陳煒群(Wei-Chun Chen)

指導教授 ： 李永銘

國立交通大學/管理學院/管理學院資訊管理學程/碩士(2020年)

摘要

本研究以Kaggle網站中預測金融詐欺的公開資料集，建立一套方法，透過SMOTE採樣方法搭配貝氏優化對XGBoost演算法調整超參數建模，預測每一筆交易其是否屬於詐欺交易，提供給銀行業建模調參的技術人員與非技術人員，以超參數調整技術（Hyperparameter tuning technologies）自動調整參數建立詐欺交易偵測模型，解決傳統建模流程超參數調優需耗費人力時間成本的痛點，以協助銀行業儘早發覺異常交易，進行風險管理。並以Kaggle 網站上的兩個信用卡交易紀錄資料集進行研究方法驗證。

關鍵字

金融詐欺 ； 超參數調整技術 ； 貝氏優化

並列摘要

This study analyzes synthetic financial datasets from Kaggle for fraud detection. The XGBoost algorithm integrated with the techniques of SMOTE sampling method and Bayesian hyperparameter optimization, is hence proposed to separate fraud transactions from non-fraud transactions. The experimental results show that our proposed method is the best predictor. Our method establishing fraud detection models for helping those people who lack the machine learning domain expertise in the banking industry detect abnormal transactions as soon as possible and carry out risk management.

並列關鍵字

Financial statement fraud ； Hyperparameter tuning technologies ； Bayesian optimization

參考文獻

[1] Lavion, Didier; et al. "PwC's Global Economic Crime and Fraud Survey 2020" (PDF). PwC.com. Retrieved 3 March 2020.

[2] H. He and E. A. Garcia, "Learning from Imbalanced Data," in IEEE Transactions on Knowledge and Data Engineering, vol. 21, no. 9, pp. 1263-1284, Sept. 2009.

[3] Chawla, Nitesh & Bowyer, Kevin & Hall, Lawrence & Kegelmeyer, W.. (2002). SMOTE: Synthetic Minority Over-sampling Technique. J. Artif. Intell. Res. (JAIR). 16. 321-357. 10.1613/jair.953.

[4] Davis, J. & Goadrich, Mark. (2020). The relationship between precision-recall and ROC Curves.

[5] CHEN, SUDUAN & YANG, ALEX. (2018). An Effective Financial Statements Fraud Detection Model. DEStech Transactions on Engineering and Technology Research. 10.12783/dtetr/pmsms2018/24902.