Credit card fraud study

With the extensive use of credit cards, fraud appears as a major issue in the credit card business. It is hard to have some figures on the impact of fraud, since companies and banks do not like to disclose the amount of losses due to frauds. At the same time, public data are scarcely available for confidentiality issues, leaving unanswered many questions about what is the best strategy. Another problem in credit-card fraud loss estimation is that we can measure the loss of only those frauds that have been detected, and it is impossible to assess the size of unreported/undetected frauds.

Fraud patterns are changing rapidly where fraud detection needs to be re-evaluated from a reactive to a proactive approach. In recent years, machine learning has gained lot of popularity in image analysis, natural language processing and speech recognition. In this regard, implementation of efficient fraud detection algorithms using machine-learning techniques is key for reducing these losses, and to assist fraud investigators.

In this paper Adaptive Machine Learning approach is utilized to detect credit card fraud.


Credit card usage is widespread in the modern societies with fraud in the credit card usage seen to be growing mostly with the recent years. The financial losses not only affect the merchants but also the banks as they have to do reimbursement whenever they have been compromised. Loss of money by banks trickles down to the customers as they are forced to pay more in form of the high interest rates that get to be charged by the banks (Siddiqi, 2013).

Get quality help now

Proficient in: Banking Frauds

5 (339)

“ KarrieWrites did such a phenomenal job on this assignment! He completed it prior to its deadline and was thorough and informative. ”

+84 relevant experts are online
Hire writer

Fraud additionally affects the reputation of the banks, financial institutions or the merchant that has been involved. This is likely to cause non-financial loss as customers will not have complete trust in such a merchant or financial service provider.

With significant threats to the credit card technology, there is need to enhance the security of such systems such that there is a possibility of fraud detection if any person would opt to commit crime in the process. Solving the issue of fraud with the credit cards is very important as several people have been in the past complained of having lost their money through fraud which if not solved could harm a significant number of people.

Loss of money through fraud usually makes the customer of a bank disappointed as they feel like the bank should ensure that the responsibility that is passed to them by their customers is enhanced. Most banks usually face withdrawal of most of their customers after suffering issues related with fraud. Basically, it would advisable for most of the credit card users to ensure that they keep their credit cards in a safe haven as this could just be yet another source through which they are to lose their critical information from fraudsters (Wheeler & Aitken, 2000).

This research paper is aimed at addressing this problem of fraudulent transactions in the credit card through analyzing an efficient fraud detection algorithm I.e Data Driven Approach which is very key in potentially reducing the conflicts. It is interesting to note that these algorithms basically rely on techniques of Advanced machine learning which will significantly help investigators in building of their investigation.

This research paper is thus going to analyze the Advanced machine learning and how it is applied to the credit card fraud detection (Wheeler & Aitken, 2000). The only shortcoming that will be faced in this research thesis is that there is very scarce public data available as most of these data is usually kept confidential as possible. Several people have lost their funds through fraud which is committed using the credit cards.

What Machine Learning is?

Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it learns for themselves. The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that we provide. The primary aim is to allow the computers learn automatically without human intervention or assistance and adjust actions accordingly.Detection of fraud in credit cards relies basically on the analysis of the transactions that have been recorded. The transaction data comprise. Automatic systems tend to be essential as human analysts are not in a position of detecting fraudulent patterns in a given dataset of transactions as the samples are large. Cardholders cannot also be relied upon to report theft or even the loss of a card as in most cases this usually several attributes which include the identifier for the credit card, the amount of transaction as well as the recipient happens with the cardholder not even being aware at some point in time. Most people usually do suffer fraud as they are never cautious with the manner in which they handle their credit cards. When using a credit card for instance in making payment say at a gas station, there is need to ensure that the details of the credit card are secure (Siddiqi, 2013). However, it is not a guarantee that the details are completely secure as fraudsters have implemented ways through which they are in a position of hacking such information. This is a problem that has been advanced by the increasing technological advancements. If it is impossible to prevent a fraud, it is effective for that particular fraud to be detected as fast as possible (Baesens, 2015). In both prevention and detection, the problem of fraud is dissected through a number of domain characteristics as well as constraints.

Problem Statement


Due to an increase in the number of fraud issues that results into loss of dollars all over the world annually Credit card fraud detection relies on the analysis of recorded transactions. Transaction data are mainly composed of a number of attributes (e.g., credit card identifier, transaction date, recipient, amount of the transaction). Automatic systems are essential since it is not always possible or easy for a human analyst to detect fraudulent patterns in transaction datasets, often characterized by many samples, many dimensions and online updates. This thesis will concern automatic data driven methods based on Machine Learning techniques. The design of a FDS based on Data Driven Models (DDMs) is not an easy task, it requires the practitioners to decide which feature to use, strategy (e.g., supervised or unsupervised), algorithm (e.g., decision trees, neural network, support vector machine), frequency of update of the model (once a year, monthly or every time new data is available), etc.Designing an efficient algorithm for fraud detection is the key for the reduction in the losses that are incurred by banks as well as other financial institutions over the years (Baesens, 2015). Several algorithms rely basically on some of the techniques of advanced machine learning so that they can assist the fraud investigators in their investigation processes. Hackers on the other hand have even advanced and gone ahead to even develop more sophisticated approaches that are used in handling of the fraud issues especially those involving the issues of credit card usage.

Purpose of the Research

The purpose of the research is to make 0% credit card fraud. In order to achieve this, we have chosen Adaptive Machine Learning methodology like Data Driven approach. With Machine Learning we let the computers to discover fraudulent patterns in the data. Data Driven approaches have some advantages, for instance with Machine Learning algorithms we can: i) learn complex fraudulent configurations(use all features available), ii) ingest large volumes of data, iii) model complex distributions, iv) predict new types of fraud (anomalies from genuine patterns) and v) adaptto changing distribution in the case of fraud evolution. With Data Driven Approach we are trying to collect huge amount of data by conducting online survey thereby we use the data for further process in order to detect the fraud.

Challenges for Data Driven

  • Unbalanced distribution
  • Concept drift due to fraud evolution and change in customer behavior.
  • For confidentiality reason, exchange of data set and information is difficult.

Research Questions

Study shall be guided by the following research questions that the researcher shall attempt to answer through the study.

  1. How is adaptive machine learning a solution to credit card fraud?
  2. How Data Driven approach is useful to solve credit card fraud?
  3. How data is collected and collected data useful for this research?


Based on the review from past studies, two main conclusions are made on the evaluation of credit card fraud detection investigations. The first conclusion is that credit card data plays essential roles in identifying fraudulent and non-fraudulent characteristics. However, the process of getting real credit card fraud related data is very hard due to record privacy and sensitivity. Therefore, to mimic the real data we will conduct an online survey and collect the user data. Based on the collected data we do manipulations with certain features that were expected to have significant impact for fraud detection. For instance, if the customer enters a wrong pin number from an actual or shipping address that was different from billing address ot transaction date or time that was too close with large sum of transactions from previous actions then it could be suspected as fraudulent.

Significance of the Study

The goal of the project is the design of mechanisms based on machine learning and big data mining techniques that allow to automatically detect attacks and fraudulent behaviorsin large amounts of transactions. Fraudulent activities are rare events that are hard to model and in constant evolution. The large volume of transactions happening everyday demands automatic tools to support investigation, and the human resources devoted to investigations have to concentrate on the most suspicious cases. This thesis investigated how machine learning algorithms could be used to address some of these issues. In particular, we focused on the design of a framework that is able to report the riskiest transactions to investigators by algorithms that can deal with unbalanced and evolving data streams. The Data Driven algorithm for a FDS depends on the accuracy measure that we want to maximize.

Ethical Consideration

Some of the ethical considerations that will be made in this research study is the aspect of data integrity and reliability. Therefore, it will be necessary for the research to observe the integrity of the data such that they do not provide data that has been cooked for the favor of the research process.

Delimitationsin fraud detection that we believe are worth investigating, such as: Defining a good performance measure, modelling Alert-Feedback Interaction and using the supervised and unsupervised information


While collecting data from a survey we may not sure that data which we have collected may change in near time. Its hard to store all the collected data and do manipulations for the expected result. Time is also a factor that needed to be considered in ensuring that the research objectives have been achieved. Limited time will be another constraint for this research study.


  1. Baesens, B. (2015). Fraud analytics using descriptive, predictive, and social network techniques: A guide to data science for fraud detection.
  2. Siddiqi, N. (2013). Credit risk scorecards: Developing and implementing intelligent credit scoring. Hoboken, N. J: Wiley.
  3. Wheeler, R. & Aitken, S. (2000). Multiple Algorithms for Fraud Detection. Knowledge-Based Systems.
  4. Zaslavsky V. &Strizhak A. (2006). Credit card fraud detection using self-organizing maps Information and Security.
  5. Bhatla, T.P., Prabhu, V., and Dua, A. (2003). understanding credit card frauds. CCardsBusiness Review# 2003-1, Tata Consultancy Services.
  6. F. N. Ogwueleka. (2011). Data mining application in credit card fraud detection system. Journal of Engineering Science and Technology, Vol. 6, No. 3 (2011) 311 – 322.
  7. Sam Maes, Karl Tuyls, Bram Vanschoenwinkel, and Bernard Manderick. Credit card fraud detection using bayesian and neural networks. In Proceedings of the 1st international naiso congress on neuro fuzzy technologies, 2002.
  8. Tej Paul Bhatla, Vikram Prabhu, and Amit Dua. Understanding credit card frauds. Cards business review, 1(6), 2003.
  9. Christopher M Bishop et al. Pattern recognition and machine learning, volume 4. Springer New York, 2006.

Cite this page

Credit card fraud study. (2022, Mar 06). Retrieved from

Let’s chat?  We're online 24/7