Financial Statement Fraud Detection Through an Integrated Machine Learning and Explainable AI Framework

Gunjargal Lkhagvadorj, Tsolmon Sodnomdavaa

Journal of risk and financial management 2025 article cited by 10

Source Open-access PDF More in Business & Industry

importance 4/10 partially solved Financial Distress and Bankruptcy Prediction financial fraud detection stage: research

Problem

Detecting financial statement fraud is hard under weak regulatory oversight, information asymmetry and severe class imbalance.

Why it matters

Undetected financial fraud harms investors and markets where oversight is weak.

What the AI did

Trained Random Forest, XGBoost, LightGBM, MLP, TabNet and a Stacking Ensemble with SMOTE and class-weighting, plus explainable AI (SHAP, LIME, counterfactuals) to detect fraud.

Results

Stacking Ensemble achieved PR-AUC = 0.93 and F1 = 0.83, outperforming baseline models.

Impact

Framework outperformed baselines on 969 firm-year observations from 132 Mongolian firms; no external adoption stated.

Record

Venue	Journal of risk and financial management
Published	2025
Work type	article
Authors	Gunjargal Lkhagvadorj, Tsolmon Sodnomdavaa
Citations	10 (OpenAlex, at import time)
References	51
Model or method	Stacking Ensemble (Random Forest, XGBoost, LightGBM, MLP, TabNet)
Organization	not stated in source
Primary source	https://doi.org/10.3390/jrfm19010013
Open access	hybrid
Keywords	fraud detection, explainable AI, stacking ensemble, SHAP, SMOTE

Importance rating

4 / 10

real but narrower advance

Real but narrow advance: fraud detection framework validated on 969 firm-year observations from 132 Mongolian firms; no broader adoption stated.

Basis: stated. The source itself states the scale. Stated scale: 969 firm-year observations from 132 Mongolian firms (2013-2024). Method and limits.

More in Business & Industry

#	Paper	Year	Citations	Importance
1	Optimizing fraud detection in financial transactions with machine learning and imbalance mitigation Ezaz Aldahasi, Rama Khaled Alsheikh, Fakhri Alam Khan, Gwanggil Jeon	2024	66	7
2	BloombergGPT: A Large Language Model for Finance Shijie Wu, Ozan İrsoy, Steven Z. Lu, Vadim Dabravolski, Mark H. Dredze...	2023	289	6
3	Search Personalization Using Machine Learning Hema Yoganarasimhan	2020	167	6
4	Combinatorial Optimization-Enriched Machine Learning to Solve the Dynamic Vehicle Routing Problem with Time Windows Léo Baty, Kai Jungel, Patrick S. Klein, Axel Parmentier, Maximilian Sc...	2024	69	6
5	A review of the applications of multi-agent reinforcement learning in smart factories Fouad Bahrpeyma, Dirk Reichelt	2022	76	6