In today's knowledge-based society, financial fraud has become a common phenomenon in most of the organizations. In order to identify these frauds, accounting and auditing firms use a technique called `Forensic Accounting', which employs accounting, auditing and investigative skills along with some mathematical models. The increasing data volume and nature of complexities in data as well as in frauds, necessitate the use of data mining techniques. This paper discusses the current systems for fraud detection; different data structures in financial accounting systems, indicating suspicious cases; and how data mining techniques can be applied to discover frauds.
In today's electronic age, financial fraud is becoming more and more commonplace. A majority of company failures can be blamed on financial factors and fraud. Fraud encompasses all irregularities and illegal acts marked by intentional deception. Fraudulently prepared financial statements are often extremely difficult for an auditor to detect. The technique, presently used by the accounting and auditing firms, to unearth manipulations of financial data is called `Forensic Accounting'. It employs accounting, auditing and investigative skills along with some mathematical models such as Benford's Law and Relative Size Factor for the purpose.
Various computer-assisted forensic auditing techniques and audit command languages assist forensic accountants in detecting manipulated discrepancies. In spite of the use of these techniques and models, the increasing data volume and nature of complexities in data, pose tremendous challenges. In this situation, data mining techniques go to the rescue of forensic accountants. Data mining is the process of discovering various trends, patterns and relationships from multiple databases and has been proved successful in recent years in various industries. This article discusses the current techniques and mathematical models used for detecting fraud in forensic accounting, wherein the "data structures of suspicious transactions, corresponding techniques to discover patterns of frauds in the database, and predictive modeling are explained". This paper presents how data mining techniques can be applied to detect fraud with an illustration.
According to Webster's dictionary, forensic accounting is an accounting method that deals with the relation and applications of the system used to record and summarize business and financial transactions, to a legal problem. It is the action of identifying, recording, settling, extracting, sorting, reporting, and verifying past financial data or other accounting activities, for evidence to be suitable for establishing accountability and/or validation of the activities of the firm. It differs greatly from regular auditing. The objective of auditing is to see whether the information disclosed by the financial statements is supported by adequate material. Auditors merely check the accounts and state that the accounts are in order after adopting due diligence. |
