Using Sparse Modeling to Detect Accounting Fraud

Author Name USUKI Teppei (KPMG AZSA LLC) / KONDO Satoshi (KPMG AZSA LLC) / SHIRAKI Kengo (KPMG AZSA LLC) / MASADA Takahiro (KPMG AZSA LLC) / SUZAKI Kosuke (KPMG AZSA LLC) / MIYAKAWA Daisuke (Hitotsubashi University)
Creation Date/NO. October 2021 21-J-049
Research Project Study Group on Corporate Finance and Firm Dynamics
Download / Links


In this paper, we implement anomaly detection on listed firms' accounting items. Using a type of sparse modeling, i.e., Graphical Lasso, we confirm that our accounting fraud detection has achieved a practically admissible level of detection capability. We also find that the method of sparse modeling contributes to detection capability.