Detecting Management Fraud in Public Companies

Document Type


Subject Area(s)

Business - Accounting


In this article, the authors describe a methodology proposed for help in detecting fraudulent financial reporting that uses basic and publicly available financial data. The methodology combines aspects of the fraud assessment research in accounting with computational methods and theory used in machine learning and datamining. The proposed approach is based on support vector machines that use a kernel that increases the power of the support vector machine because it is developed with a domain specific to finance. The results of using the methodology are presented.

Rights © 2010 by INFORMS (Institute for Operations Research and Management Sciences)

This document is currently not available here.