details
MESA: A New Methodology for Efficient Detection of Fraud in Financial Statements
01-01-2015 to 31-12-2017
The purpose of the project is to increase the effectiveness of fraudulent financial statement detection. Until now, the applied use of software to support the detection of fraud in financial statements has been limited due to the clumsy input task required, the poor predictive power obtained, and the “black box” character of models employed in prediction. The project has surmounted such difficulties. First, a new theoretical basis for the modeling of multivariate relationships where accounting information plays a role is deduced from simple premises. Then, from such foundation, the project has built a decision support system (the MESA DSS), able to increase significantly the predictive power of financial statement fraud detection while using interpretable models. Web-mining is also utilized to automate data input. A “beta” version of the MESA DSS software is at http://accounting-analytics.net/
Research leading to the MESA DSS and to related theoretical developments is funded by the Foundation for the Development of Science and Technology (FDCT), a public institution of the Macau SAR of China.
