History of credit card goes back to early 1900s but first card was issued by banker in Brooklyn title as charge-it in 1946. But, it all began when Mr Frank McNamara and his partner Ralph Schneider paid their restaurant bill with cardboard card coined as Diner’s club card in the mid 1950s. It was the first credit card in widespread used in history (Gerson & Woolsey, 2009). Back then probably even they must haven’t imagined that after more than half century, 686 million credit cards in US and in UK about 80 million debit card and about 60.7 million credit cards will be in circulation (Woolsey & Schilz, 2011).
If you need assistance with writing your essay, our professional essay writing service is here to help!Essay Writing Service
With this abundance of the card and transactions, there is another fact that from 2006 to 2008, UK alone has lost £427.0 million to £609.90 million due to credit and debit card fraud. And study shows that in 2006, cost of fraud was 7 cent in every $100 transaction in US (Woolsey & Schilz, 2011) (Financial Fraud Action UK, 2011). This shows that how the financial world is suffering from the fraud and theft every year and ultimately public have to pay the price of such financial crime.
Thus, to detect, subdue and avert these frauds, banks and financial institutes researching and developing better ways from years. Credit cards have changed from cardboard card to plastic cards, hand-signature to digital magnetic strip to embedded chips. Traditional back room accountant have upgraded into digital computing machines. Complex statistical and analysis algorithm of data mining have evolved into artificial intelligence in expert system, but still both fraud techniques and fraudster have been evolving forcing banks and financial institutions to evolve their ways for detection and prevention together.
Fraud is human behaviour of intentional deception for gaining advantages of financial or other aspects. In later years, credit card fraud has grown to a serious and threatening financial crime. This crime involves counterfeit or hijack of one or several legitimate attributes of the user, by means of identity theft, physical theft of cards, hacking of online user credentials etc, and using such illegitimate information for own advantages. Data mining is one of the effective methods of controlling credit card fraud.
Data mining is technique of extracting unseen and previously unutilized information from the available data. At present, fraud analysis and detection have received major attention due to its capabilities of extracting hidden information and using it to efficiently classifying or predicting new information. Many methods and algorithms are employed for data mining, from plain statistical analysis to sophisticated artificial intelligence.
Artificial neural network, a computing model analogous to biological neurons of brain, is quite popular for developing intelligent agents for mining data. Artificial neural network is collection of simple processing elements, units or nodes, interconnected for processing using connectionist approach to computation or parallel distribution processing model. Its structure is adaptive; it changes its structure to internal and external flow of information. These networks are mostly used to form links from inputs to outputs or to extract patterns in data.
Credit card fraud stands as major problem for word wide financial institutions. Annual lost due to it scales to billions of dollars. We can observe this from many financial reports. Such as (Bhattacharyya et al., 2011) 10th annual online fraud report by CyberSource shows that estimated loss due to online fraud is $4 billion for 2008 which is 11% increase than $3.6 billion loss in 2007and in 2006, fraud in United Kingdom alone was estimated to be £535 million in 2007 and now costing around 13.9 billion a year (Mahdi et al., 2010). From 2006 to 2008, UK alone has lost £427.0 million to £609.90 million due to credit and debit card fraud (Woolsey & Schilz, 2011). Although, there is some decrease in such losses after implementation of detection and prevention systems by government and bank, card-not-present fraud losses are increasing at higher rate due to online transactions. Worst thing is it is still increasing un-protective and un-detective way.
Over the year, government and banks have implemented some steps to subdue these frauds but along with the evolution of fraud detection and control methods, perpetrators are also evolving their methods and practices to avoid detection. Thus an effective and innovative methods need to be develop which will evolve accordingly to the need.
Nature has always been inspiration of the scientific studies and experimentation. One of the most mysterious challenges for the scientist is working of the human brain. Imitating the human neuron, computer scientist has developed a computing model artificial neural network. For the fraud detection and developing expert system, ANN is widely used but still some research have shown that NN may not be right choice for the detection as in research other techniques have outperform NN (CCFD Bayesian and NN)(comparison of classification method applied on CCFD). But they have suggested to optimization and pre-processing of data can increase the efficiency of ANN.
The main objective of this project is to observe and implement the optimization method for the artificial neural network and compare and evaluate the result. Since the use of ANN is mainly in developing artificial intelligence and expert system, optimization is necessity. Thus our main aim in this project is to find out the efficient method of optimization of neural network for classification of good and bad credit transactions.
There are two types of methodologies to follow for research, quantitative and qualitative. As for this project, we have to deal on collection of simulation data and perform analysis; we will follow the qualitative methodologies. We will follow RAD (rapid application development) model for building our model of neural network. We will feed the test data set to our model and calculate efficiency of the network against different optimization method. Later we will analysis the efficiency of each model and make our suggestions.
Literature Review of related work
In a book of data mining, (Witten & Frank, 2005) defines data mining as the process of discovering patterns in data, automatically of semi-automatically, which is meaningful and will lead to some advantages preferably economical. Book further adds, data mining is about solving problem by analyzing already present data in database. (Palace, 1996) Generalize data mining or knowledge discovery as the process of analyzing data from different perspective and summarizing them into useful information finding correlations or patterns in many fields in large relational database.
Our academic experts are ready and waiting to assist with any writing project you may have. From simple essay plans, through to full dissertations, you can guarantee we have a service perfectly matched to your needs.View our services
In article, (Fayyad et al., 1996) and (Palace, 1996) classifies usages of data mining in mainly four tasks, clustering, classification, association rule learning and regression which are used independently and/or in combined form in many fields such as business and marketing, scientific study & technology, security and surveillance and pattern mining in spatial data. In their paper, (Phua et al., 2010), they have briefly explained data mining and fraud. According to them, data mining as about finding insights which are statistically reliable and unknown earlier, and tools to solve problem which can’t be done by query and reporting tools. Also fraud as business critical problem and its detection and prevention is imperative in any competitive environment.
In article, (Palishikar, 2002) author recognizes fraud as white colour crime responsible to bleed hundreds of billions from business from worldwide. They also identify, data mining as a help to mine evidence of fraud and fraudulent activities from the huge amount of data. (Mahdi et al., 2010) Agree in fact that fraud and credit card fraud instances can be similar in content and appearance but they usually are different. They come in different shape and sizes. Thus it needs complex and time consuming investigations (Palishikar, 2002). Recognizing the potential, telephony companies were among the first to use data analysis techniques to prevent fraud. (KUÅÅ¾AKSIZOÄÅ¾LU, 2006).
PhD. thesis, (Paasch, 2008) explains extensively about credit card transaction and its various fraud types. (Paasch, 2008) and (Bhattacharyya et al., 2011) divides statistical fraud detection methodologies in two approaches, supervised and unsupervised. The paper, (Bhattacharyya et al., 2011) further classifies the fraud detection methodologies into three data mining technique logistic regression (LR), support vector machines (SVM), random forest (RF). In paper, (Srivastava et al., 2008) authors have suggested that credit card fraud detection have attracted many research interest over many techniques. Paper further makes special emphasis on data mining and use of neural network.
From the early 90’s to late 2010 there are number of other account of researches that are dedicated for credit card fraud detection. Some of them are purely sequence alignment algorithmic analysis such as, (Kundu et al., 2009) and (Srivastava et al., 2008). Some are purely sophisticated system as mention by (Aleskerov et al., 1997) Clementine, Darwin, Falcon and PRISM. They are based on various data mining techniques such as classification, visualization, segmentation, clustering, profiling, deviation detection, and association rule generation, K-Nearest Neighbour (KNN) method implemented independently or combinational form in building Neural network, decision trees, fuzzy logic and expert systems etc. further it report of resulting in 70% fraud detection rate.
According to (Bhattacharyya et al., 2011), both supervised and unsupervised fraud detection methods are based on predictive model. Paper also accents on active use of predictive models of credit card fraud detection. It further reports, on relatively few studies conducted on this topic. Among those few studies, primary focus is upon neural network and its variance of method of applications (Maes et al., 1993) (O`Dea et al., 2001) (Gupta & Bhargava, n.d.) (Aleskerov et al., 1997) (Zhu et al., 2010) (R. Brause, n.d.).
In research paper published back in 1994, (Ghosh & Reilly, 1994) researchers had proposed use of neural network for detecting credit card fraud. Researchers had built a system to detect by training on large sample of labelled credit card account transactions which contains fraud cases due to lost cards, stolen cards, application and counterfeit fraud etc.
Another complete system builds for same purpose using meta-learning techniques to learn model of fraudulent credit card transactions (Stolfo et al., 1997). Recent research accounts of development of fraud detection system using PGNNs (parallel granular neural networks), for increasing data mining and knowledge discovery process (Syeda, 2002).
Artificial neural network, a popular term since 90s and mention several time in above literatures, is defined as “A neural network is an interconnected assembly of simple processing elements, units or nodes, whose functionality is loosely based on the animal neurons” (Gurney, 1997) and neural network when used for & with intelligence then is called Artificial Neural Network (Gupta & Bhargava, n.d.). With definition, authors have explained about working of ANN in the paper. In (Gupta & Bhargava, n.d.), various type of architecture of ANN is explained such as Kohonen’s self organizing maps and simple Hopfield network. This paper concerns application of ANN in business applications, mainly in predicating anomalies, fraud, bankruptcy etc.
There are number of research paper which explains the use of neural network for specific credit card fraud detection. As many of researches there are as many variances of the approached. Such as use of Bayesian network and neural network (Maes et al., 1993), use of combined probabilistic and neuro-adaptive approach (R. Brause, n.d.), feature selection approach before classifying using neural network (O`Dea et al., 2001), PGNN for fast learning and detection (Syeda, 2002), optimization using Genetic Algorithm for Artificial Neural Network (Paasch, 2008), and study of use of chaos theory and neural network for credit card risk detection (Zhu et al., 2010). From above literatures we can observe many researches are done and many are ongoing for developing optimized, efficient and cost effective method to detect risk and fraud in direct or online transaction involving credit / debit cards which are based on data mining and neural network.
Cite This Work
To export a reference to this article please select a referencing stye below:
Related ServicesView all
DMCA / Removal Request
If you are the original writer of this essay and no longer wish to have your work published on UKEssays.com then please: