Attached.

Running Head: BANK FRAUD DETECTION

1

Bank Fraud Detection: Support Vector Machine Model
Name
Institution

BANK FRAUD DETECTION

2

The significant development and advancement in computing and communication have
resulted in increased cases of fraud in terms of the forms and amount of fraud. According to
Abdelhamid et al. (2014), fraud refers to various forms of tricks used by criminals to gain
financially. Financial institutions such as banks operate under strategic goals, and objectives.
However, the mission is to attract and retain new and existing customers by ensuring that they
receive value. However, fraudulent activities make it difficult to achieve these goals due to a lack
of trust. The targeted bank, XYZ, is vulnerable to fraud, but credit-based fraud is the most
common and at risk of happening. In this case, fraud detection and prevention is the most
effective method of protecting banks to ensure they offer satisfactory services to their customers.
Fraud detection goes beyond prevention since it comes into play when fraudulent transactions
start happening.
The XYZ bank generates large volumes of data, including customer profiles, transaction
history, among others are collected daily, and this data can play a critical role in ensuring
suspicious activities are detected to prevent fraud. There exist many methods of fraud detection
and prevention. However, deep machine learning is considered to be the most effective in the
modern world of technology. This involves the use of big data applications and data mining
techniques to make fraud prediction. Data mining involves the extraction of crucial information
from a large volume of data (big data), which can help in the decision-making process. In this
paper, we will analyze how machine learning and data mining tools and techniques help in
ensuring early detection of fraud from data generated in the bank.
Problem Statement
The financial service sector is currently going transformation due to the increased
adoption of AI technology. This technology has been utilized in various segments, including the

BANK FRAUD DETECTION

3

management of the portfolio, customer service, improving customer experience, increasing
regulatory compliance, and assigning credit scores, among other uses. In this case, tasks that
traditionally required a large number of workers and hours to achieve can be done in seconds.
However, fraud in the financial sector and especially banks have increased, and Anan et al.
(2018), provides that cybercrime, which constitutes financial fraud adds up to over $600 billion
globally, which is almost one percent of the global GDP. The problem facing the Banking
industry and the target bank is that cybercrime has increased, and cybercriminals are becoming
smarter as technology develops. This means that the bank does not have a choice but to improve
its defense and develop prevention and detection capabilities much faster.
There exist different forms of bank fraud, most of which have evolved from increased
threats by cybercriminals. For instance, money laundering is the most common type of fraud.
This is an illegal activity whose aim is to hide the source of money by carrying out a sequence of
complex banking transactions to avoid detection. Many states have put in place mechanisms to
detect this kind of transaction and prosecute those involved. In this fight, investigations involve
the analysis of suspicious transactions by the banks (Gyamfi & Abdulai, 2018).
The other common form of fraud is credit card fraud. This takes various forms and
involves the unauthorized use of a stolen credit card to extort money or obtain services. A report
by Anan et al. (2018), shows that global losses arising from credit card fraud could reach $44
billion by 2025. Different methods can be used to detect credit card fraud, most of which involve
the use of forecast indicators retrieved from transaction information. Other forms of banking
fraud include identity theft, mobile theft, among others. As a result of these fraudulent activities,
banks are suffering due to direct costs, loss of custo...