Data Mining and Accounting Fraud

Save this page for later by
adding it to your bookmarks
Press Ctrl+D (Windows)
or Cmd+D (Mac OS)
Text
Sources

Abstract

With increased financial accounting fraud been experienced in the contemporary economic situation, financial accounting detecting fraud (FADF) has turned out to be a significant topic in research and academic. Organizations are now using specialized methods to detect accounting fraud (largely known as forensic accounting) because internal auditing systems have failed to identify accounting frauds. In accounting fraud detection, data mining techniques are providing enough support considering dealing with complexities of financial data and large data volumes are big challenges for forensic accounting. In this accord, this paper offers an extensive literature review on application of data mining methods for the detection of accounting fraud. An analysis indicate that data mining methods such as decision trees, neural networks, logistic models, and Bayesian belief network have been used comprehensively to offer first hand solutions to problems intrinsic in detection and classification of fraudulent information.

Executive summary

With increased financial accounting fraud been experienced in the existing economic situation, financial accounting detecting fraud (FADF) has turned out to be a significant topic in research and academic. Organizations are now using specialized methods to perceive accounting fraud (largely known as forensic accounting) because internal auditing systems have failed to identify accounting frauds. In accounting fraud detection, data mining techniques are providing enough support considering dealing with complexities of financial data and large data volumes are big challenges for forensic accounting. In this accord, this paper offers an extensive literature review on application of data mining methods for the detection of accounting fraud. An analysis indicate that data mining methods such as decision trees, neural networks, logistic models, and Bayesian belief network have been used comprehensively to offer first hand solutions to problems intrinsic in detection and classification of fraudulent information.

turnitin
We can write
your paper for you
100% original
24/7 service
50+ subjects

Introduction

With increased financial accounting fraud been experienced in the existing economic situation, financial accounting detecting fraud (FADF) has turned out to be a significant topic in research, academic and has grabbed increased attention from investors, media, regulators, and financial community. Because of some high profile accounting frauds exposed and reported in large firms like WorldCom, Enron, Satyam, and Lucent over the last ten years, the need of defining, perceiving, and reporting accounting fraud has augmented.

Fraud according to the Oxford English Dictionary (2015) is a ‘criminal or wrongful sham anticipated to lead to personal or financial gain.’ But, in theoretical literature, fraud is demarcated as resulting to the exploitation of a profit organization’s system devoid of essentially leading to undeviating legal concerns (Phua et al., 2005). Even though this literature seems to miss a collectively acknowledged meaning of accounting fraud, Wang (2010) demarcated it as a ‘measured act that is opposing to rule, law, or policy with intent to obtain illegal financial benefit.’ Hence, by making forged financial accounting statements, where numbers are influenced by understating debts, expenses, or liabilities, spurious entries related to profit and sales, overstating assets, and or misappropriation in taxes, accounting fraud is executed.

Accounting fraud is turning out to be an increasingly serious problem economically but effective detecting of accounting frauds remains a complex and significant responsibility of accounting experts. In organizations, internal auditing has become a challenging activity and there exists proof that ‘book-cooking’ accounting practices are being used for conducting accounting frauds while traditional internal audit procedures are turning out to be difficult in detecting accounting frauds. In the current competitive business environment, without guaranteed accounting fraud prevention procedures and tools, accounting fraud has turned out to be a business critical problem. However, the entire process is automated by data mining techniques which also help lessen the manual work of checking and screening several statements

Data mining techniques are providing enough support considering dealing with complexities of financial data and large data volumes are immense challenges for forensic accounting. In this accord, this paper offers an extensive literature review on use of data mining methods for the uncovering of accounting fraud. An analysis indicate that data mining methods such as decision trees, neural networks, logistic models, and Bayesian belief network have been used comprehensively to offer first hand elucidations to problems intrinsic in uncovering and cataloguing of deceitful information.

Literature review

There exists many examples firms which have suffered because of fraud committed by accounting and other interrelated departments. This has resulted to several researches been conducted on the subject. Previously, studies on accounting fraud primarily focused on fraud risk factors (such as red flags). This red flags helped researchers determine factors which are connected to sham financial activities. As noted, accounting fraud occurs when financial statements are altered either by using fictitious records or recording revenues prematurely. Nonetheless, in finding financial statement fraud, the familiarity of auditors is considered a significant factor (Sharma and Panigrahi, 2012). Olszewski (2014) for instance established that the experience of auditors was an important factor. Knapp and Knapp (2010) correspondingly examined the impacts of risk assessment and audit experience and found out that audit managers are more real in measuring fraud risk.

Investigative techniques have been used widely to detect fraudulent accounting practices. Several studies have previously focused on non-fraud and fraud companies and established that capital turnover, firm size, financial leverage, and asset composition were important factors which impacted the possibility of fraudulent accounting reporting. New models recently are been relied on to detect fraudulent accounting practices. Huang et al. (2014) for example premeditated the impact of using expert systems on the performance of auditors in identifying fraud. The study established that using expert systems auditors are more consistent in their decision an in a better position of detecting accounting fraud. Based on neural networks (one of data mining techniques), Green and Choi (1997) developed a model. The outcomes of their study indicated that neural networks are vital in detecting accounting fraud.

To predict financial distress, Sun and Lee (2006) used data mining methods by combining decision tree model, information gain, and attribute-oriented induction in their study using data from different companies and different financial ratios. This study found that data mining methods are valid and feasible for perceiving accounting fraud. Also, a study by Kirkos et al. (2007) used data mining techniques for fraud detection. However, this study compared more than a few methods by their performances. Chen and Du (2009) in a recent study used artificial neural networks and data mining techniques using different firms in Taiwan Stock Exchange. According to the results of the study, in predicting financial distress of the companies, artificial neural networks were much better than outdated statistical methods. All these studies indicate that data mining methods are becoming quite important in detecting accounting fraud.

Deadlines from 1 hour
Get A+ help
with any paper

Analysis

Organizations in contemporary business world are in better position to advantage from the information they gather from their partners, customers, environment, and processes especially with the introduction of information technology which facilitates and enables the collection, processing, and storage of large amounts of data. However, the massive sophistication and amount of information gathered needs the use of innovative techniques such as data mining to extract definition from the raw data and use it for tenacities advantageous to the company. According to Sahin et al. (2013), data mining defines the analysis of information to ascertain formerly unidentified relationships that offer useful information. To elucidate problems and expand the several facets of business, data mining has been embraced by several industries like telecommunication, retail, healthcare, and finance.

Basically, data mining functions by examining information, and producing predictive and descriptive prototypes which aid to solve problems. In practice, several classifications of data mining applications are used. It has been classified into six categories by Larose (2005) including prediction, estimation, association, description, clustering, and classification. Fraud detection falls under the classification category. For the mentioned classes of problems, several techniques have been developed such as neural networks and decision trees. However, it is vital to note that data mining is a method with more than a few stages during which cautious human interference and understanding are required even though there exists many commercially accessible software packages with comprehensible graphical borders making complex data mining tasks apparently minor to apply.

Recommendation

To detect accounting fraud effectively and be able to prevent it, it is important for auditors and company financiers to focus also on other important areas of finance and accounting (like auditors statements and unaudited financial information), not only on financial statements data. It is also important to embrace data mining techniques in fraud detection as the benefits of using them is comprehensive.

Need a custom paper ASAP?
We can do it today.
Tailored to your instructions. 0% plagiarism.

Summary and conclusions

The paper reviewed the literature on the relationship between data mining and accounting fraud and provided an analysis describing data mining techniques in detail. The study established that many companies and industries are using regression analysis to detect fraud because it has a better explanation ability. However, in the case of data mining, it seems neural networks are quite important in fraud detection. Even though there exists no comparison between accuracy of neural network and detecting effect, and regression model, neural network seems to have many advantages. Organizations are now using specialized methods to detect accounting fraud (largely known as forensic accounting) because internal auditing systems have failed to identify accounting frauds. In accounting fraud detection, data mining techniques are providing enough support considering dealing with complexities of financial data and large data volumes are big challenges for forensic accounting.

Did you like this sample?
  1. Chen, W.S & Du, Y. K (2009). Using Neural Networks and Data Mining Techniques for the Financial Distress Prediction Model. Expert Systems with Applications, Vol. 36, pp. 4075–4086
  2. Green, B. P & Choi, J. H (1997). Assessing the Risk of Management Fraud through Neural Network Technology. Auditing: A Journal of Practice and Theory, Vol.16, No.1, pp.14–28
  3. Huang, S. Y., Tsaih, R. H & Yu, F (2014). Topological Pattern Discovery and Feature Extraction for Fraudulent Financial Reporting. Expert Systems with Applications. 41, 4360–4372.
  4. Kirkos, E., Spathis, C & Manolopoulos, Y (2007). Data Mining Techniques for the Detection of Fraudulent Financial Statements. Expert Systems with Applications, Vol. 32, No.4, pp. 995–1003.
  5. Knapp, C. A & Knapp, M. C (2001). The Effects of Experience and Explicit Fraud Risk Assessment in Detecting Fraud with Analytical Procedure. Accounting, Organizations and Society, Vol. 26, pp. 25-37.
  6. Larose, D. T (2005). Discovering Knowledge in Data: An Introduction to Data Mining. John Wiley & Sons, Inc., Hoboken, New Jersey. pp. 11-17
  7. Olszewski, D (2014). Fraud Detection Using Self-Organizing Map Visualizing the User Profiles. Knowledge-Based Systems. 70, 324–334
  8. Oxford Concise English Dictionary (2015). 11th Edition. Oxford University Press.
  9. Phua, C., Lee, V., Smith, K. & Gayler, R. (2005). A comprehensive survey of data mining-based fraud detection research. Artificial Intelligence Review 1–14.
  10. Sahin, Y., Bulkan, S & Duman, E (2013). A Cost-Sensitive Decision Tree Approach for Fraud Detection. Expert Systems with Applications. 40(15), 5916–5923
  11. Wang, S. (2010). A Comprehensive Survey of Data Mining-Based Accounting-Fraud Detection Research. International Conference on Intelligent Computation Technology and Automation, vol. 1, pp.50-53.
Find more samples:
Related topics
More samples
Related Essays