Neural Networks and Fraud Prevention (Essay Sample)

Name:
Institution:
Instructor:
Course:
Date:
Neural Networks and Fraud Prevention
Introduction
Fraud in the business environment is increasing in each and every year and it is regarded as a million dollar business. It is alleged that both external and internal forms of frauds presents substantial costs in the economy of various countries. The review of most academic literature indicates that most academicians have only examined the causes of internal frauds and means of prevention and detection. In this case, therefore, little effort has been put in place to investigate on how to detect and prevent internal fraud, which is otherwise referred to as internal fraud risk reduction’. This essay presents the internal and external framework which can be deduced from both the current business practices and academic literature, where the core of the framework suggests the usage of the data mining approaches. Arguably, internal fraud is a major problem in the economy of most countries. Most organizations have allocated lots and lots of resources to internal control. Arguably, this is a framework employed by most business organizations in businesses in preventing internal frauds. This kind of costs as well as the costs of internal frauds represents large economic costs in the business environment if it goes unnoticed.
What is Fraud?
Fraud can be defined as some form of deception. It should be noted that whatever the industry is fraud situated or whatever form of fraud that one can visualize, the main core of fraud is always deception. There are various definitions which have been given to fraud. The American Heritage Dictionary defines fraud as “a deliberate deception that is practiced so as to secure unlawful or rather unfair financial gains” (Abidogum, 205). Daviaet al. 2000, on the other hand, argues that fraud tend to involves an individual or more than one person who have the intention of depriving another person a valuable thing for their own financial enrichment.
Classification of Fraud
In this case, the delineation of fraud to occupational abuse and fraud is one of the various ways in which fraud can be classified a classification resembling this delineation is the distinction made between external and internal fraud by Bologna and Lindquist (1995) notably, this type of classification is mainly applied organizational settings, and it is based on the assertions that perpetrators are external or internal to the victim companies. Arguably, frauds which are committed by suppliers, vendors and contractors form some of the examples of the external types of frauds. On the other hand, employees who steal from the company by manipulating accounting books are some examples of the external frauds. Other ways through which fraud can be classified is known as statement versus transaction fraud. In this case, statement fraud is defined by Bologna and Lindquist (1995) as “the intentional misstatements of financial values to enhance appearances of profitability hence deceiving creditors or shareholders”. Accordingly, the main aim of transaction frauds is to embezzle assets of an organization. Assets- theft fraud and financial statements balance fraud are two types of frauds which tend to be related. According to Daviaetal(2000), the major differences between financial statement balance asset and the other type of fraud is that theft of assets is not involved in the financial statement fraud
Costs of Fraud:
As already noted, fraud is a million dollars business. According to the survey conducted by the PricewaterCoopers in 2007, 43 percent of the companies surveyed across the world have fallen victim to fraud in the years 2005 to 2007. The average costs to these companies were estimated to be more than US$ 2.4 millions for every company. Arguably, there is no company or industry that is safe from fraud, and larger companies tend to be the most vulnerable as compared to the smaller companies. During this survey, it was established that at least 5% of the company’s total annual revenues is lost to fraud. If applied to 2006 US’Ss GDP, then this is approximately US$ 660 billions in fraud loss for the US only. It should be noted that the figures mentioned above are all as a result of internal frauds. This notwithstanding, however, there is also significant costs which results from external frauds. Accordingly, the major domains that are regularly inflicted by fraud include health care, automobile insurance, telecommunications and credit cards. Globally, fraud in the telecommunication industry is estimated to be US$ 50 billions, while the costs of automobiles insurance frauds is estimated to be US$ 25 billion in the United States.
Detection versus Prevention
There is several ways of detecting fraud. However, some authors such as Lindquist and Bologna (1995) argue that prevention of fraud should be given priority as opposed to detection. In this case, the two authors allege that fraud can be prevented by creating a working environment that values honesty. This will include the hiring of employees who are honest in their work, treating employees fairly and paying them well. Moreover, employees have the mandate of providing a secure and safe workplace. According to the Accountant's Guide, it is the responsibility of the management to allocate enough resources, and also to put emphasis on “fraud that is specific to internal control and to proactive examinations which are fraud-specific” (Daviaet al, 35). This kind of approaches is an example of detection in one side, and prevention on the other.
NEURAL NETWORKS AND FRAUD DETECTION
How neural networks are being used to prevent fraud by data mining, by tracking inconsistencies in transaction activities for payment transactions for online consumer businesses, or by banking institutions. The modern technology is expanding daily in every part of the world. This has improved the communication systems which has benefitted many especially the business entrepreneurs. With the many advantages of the modern technology, fraud has dramatically increased. As a result many businesses have lost billions of dollars mysteriously. Prevention technologies have been established as the best way to tackle fraud but fraudsters have with time found their way through. Some of the fraud activities most fraudsters have indulged in are money laundering, e-commerce credit card scam, telecommunication frauds well as computer intrusion.
The neural network is an information processing model controlled by the way the nervous system such as the brain receive and synthesizeinformation. The important elements to the artificial neural network are the neurons. They are highly interconnected processing elements that work together to tackle a similar problem. The development of the neural networks was before the advent of the computers. They have been useful since they are able to extract patterns as well as trends that are difficult to be noticed by humans.
Today, neural networks have been largely put in use to prevent fraud in the banking industry. During payment of salaries in the banks fraud has evolved constantly. The fraudsters are always on the lookout for any loopholes in the payment system so as to take advantage. They seek to maximize on the results of their activities. Those who offer payment services, issuers,banks as well as merchants have adopted neural networks as the main tool to prevent fraud. Fraud detection is a process that is done in the banks that enables the separation of transactions that are vulnerable to fraud and those not. The patterns in the data are used to do this. The Bayesian models together with the neural networks are used in different ways for fraud detection.
Fraud detection is not an easy task. Moreover responding to a fraudulent occurrence is not easy. The fraud detection systems are required to contain highly skewed allocations of figures. This is because it is only a small percentage of the transactions are fraudulent. Another difficulty in fraud detection is the presence of errors in the data and poor maintenance of the record. Those with errors are supposed to be filtered out and the records managed in such a way that the transactions assumed to be legitimate are minimized whereas they are fraudulent. The detection system has to be able to adapt to different frauds. If they fail frauds are able to go through undetected.
Data mining techniques have been widely put in useto prevent and detect financial frauds. The implementation of the techniques to detect fraud has to follow the traditional information. This is the flow of data mining; which starts with selection of feature selection, representation, data collection and management, pre - processing, data mining, post-processing, and finally performance evaluation. Data mining techniques succeed in detecting fraud because they use past circumstances of fraud to form models to be used to detect the jeopardy of fraud.
Financial statement fraud is one of the main financial frauds that is rampant worldwide and it has caused big companies to collapse due to financial losses. This has left a bad picture on the efficiency of corporate governance as well as the quality and credibility of financial reports. This fraud of financial statement fraud is a serious issue in the businesses globally.
A neural network to detect fraud detection is an essentialapplication of Data Mining. Both researchers’ and practitioners have accepted that the analytical procedures, data mining techniques along with traditional reviewing procedures are necessary to prevent and detect financial statement fraud.
ATM AND CREDIT CARD FRAUD DETECTION
In the recent years as more people adopt the use of credit cards and ATM to carry out their transactions fraud has as well taken root. Difficulty has been i...