Examining the History of Fraud

Within Banking and How Advanced

Technology Plays a Role

Gabriella Villacampa / Intro to Business Analytics / Student ID Number: 13921985

Table of Contents
Part 1: Wells Fargo’s Use of Business Analytics to Overcome Fraud
Introduction……………………………………………………………………………....4
Use of Data Analytics Within Banking………………………………………………....5
Two Main Types of Data Analysis:…………………………………………………………….....5
i. Operational Analysis……………………………………………………………….....5
ii. Strategic Analysis……………………………………………………………………..5
Data Analytics in Wells Fargo:…………………………………………………………………...5
i. Centralized Data……………………………………………………………………...5
ii. The Big Idea…………………………………………………………………………..6
Machine Learning………………………………………………………………………..6
i. Cyber Security:…………………………………………………………………….......6
a. Data Integrity……………………………………………………………………...7
b. Security Awareness Training……………………………………………………..7
ii. Blockchain:…………………………………………………………………………....7
a. Fraud Reduction…………………………………………………………………..7
b. Know Your Customer (KYC)…………………………………………………….8
The Process of Predictive Analytics………………………………………………….....8
Methods of Predictive Analysis:…………………………………………………………………..8
i. Data Mining…………………………………………………………………………..8
a. Transaction Monitoring…………………………………………………………..8
b. Account Profiling…………………………………………………………………9
ii. Regression Analysis…………………………………………………………………..9
a. Risk Scoring………………………………………………………………………9
b. Variable Selection………………………………………………………………...9

2
Table of Contents
Part 2: Credit Card Approvals
Introduction……………………………………………………………………………..10
Descriptive Statistics Analysis of Selected Variables…………………………………11
Hypothesis……………………………………………………………………………….12
Hypothesis A……………………………………………………………………………..12
Hypothesis B……………………………………………………………………………..14
Recommendations for Enhancing Analytical Skills to Better Combat Fraud Within
Wells Fargo……………………………………………………………………………...16
Recommendations for Improving the Analysis Within the Dataset………………....17
Conclusion………………………………………………………………………………18
Bibliography…………………………………………………………………………….19

3
 Part 1: Wells Fargo’s Use of Business Analytics to Overcome Fraud
Introduction
Fraud is as old as money itself and examiners have gone a step further to say that fraud
has most likely existed for as long as humans have inhabited this world. For many years, banks
have fallen victim to fraud and LexisNexis ran a study named “True Cost of Fraud” where they
held a survey with 501 risk executives across the U.S. and Canada. The study found that fraud
costs up to 22.4% from Pre-Pandemic levels across these financial services. “Based on attacks
and costs, the costs continued to rise above early 2020 levels, with banks reporting the highest
figure of $4.36 for every $1 of fraud loss” as stated by LexisNexis (2022). 36% of these losses
came from new accounts and there were also increased investments into risk mitigating solutions
such as digital identity.
With banks being an easy target for fraud, they have looked into ways to improve their
technology so that they can provide advanced solutions when it comes to preventing it. Wells
Fargo, for example, is proactively enhancing their security to recognize new threats and is using
multiple layers of protection to improve the safety of accounts and information. As mentioned on
the Wells Fargo (no date) website, “as fraud becomes more sophisticated, so does our approach
to security – and our investment in advanced security tools using the latest technology helps us
protect your accounts and information and detect when there may be a risk”.
Business Analytics continues to play a major role in combatting fraud losses through
techniques and, according to Gupta et al (2019), “analytics is basically the term which we use to
define the process of taking in the data and after the execution, apply some mathematical or
statistical techniques or even some tools to determine the useful information from the
information stored”. Banks that can support advances in technology and analytics to refine fraud
prevention will have a better chance in increasing their growth rate. Based on the statistics
examined by Gupta et al (2019), “12% of banks are analyzing big data, 25% of banks are
increasing their growth rate by processing it and 38% of banks are expanding various paths”.
There are some considerate steps that can be taken in order to control fraud such as education on
fraud prevention, increasing the power of data analysis technologies and displaying the best
practices for fraud reduction. Even though banks cannot be 100% sheltered from unknown
threats, an undeniable level of preparedness can go a long way in countering fraud risk.

4
Use of Data Analytics Within Banking
Bănărescu (2015) describes data analysis as “an in-depth examination of the meaning and
essential features of available data in order to identify significant information using specific
methods and techniques”. This examination can be very tedious as its goal is to propose guiding
lines by using the data it analyzes to recognize data gaps, strengths, vulnerabilities and risk
factors that can then form into threats. Data analysis technologies permits auditors and fraud
examiners to inspect many businesses’ forms of data to identify how their internal controls are
operating and pinpoint transactions that stipulate any fraudulent identity or a high risk of fraud.
According to Revathi Subramanian (2014, pg. 22), “Getting fraud data accurately captured,
categorized and stored is the first important step to using data-driven technology to combat fraud
losses”.
Two Main Types of Data Analysis:
i. Operational Analysis: Data analysis, in operational form, develops into a useful tool
to revamp the workplace conditions, minimizing labor activities and mental abilities
when it comes to preparing large amounts of data. Operational analysis can be
strongly used in the short-term by utilizing data and current information as a ground
to follow the present activities in order to detect fraud with intense planning. As
Bănărescu (2015) states, “the main role of operational analysis is to help antifraud
managers to detect and combat illegal activities by examining (1) links between
suspects, (2) their characteristics (key positions that impact decision-making, etc.), (3)
the movement of goods, money, or other valuables, (4) way of communication and
(5) sequence of certain events”.
ii. Strategic Analysis: Unlike operational analysis, strategic analysis requires a larger
scale approach when it comes to combating and identifying fraud. The analyst works
with large volumes of data by using statistics to formulate explanations and
predictions in order to corroborate decisions for the highest level of management
(Bănărescu, 2015). When using digital statistical tools, someone should have the
ability to spot strange disparities by looking at the curve. “With the new generation of
visualization software, we can dive into massive data sets and visually find new
trends, patterns and threats that would take hours or days using conventional data
mining” (Bresfelean et. al, 2008).
Data Analytics in Wells Fargo:
i. Centralized Data: Enterprise analytics was introduced to the company to provide a
better understanding and with the advancement and improvement of AI and machine
learning, Wells Fargo had a goal to create a center of excellence in order to bring the
latest and greatest into the bank. As explained by Prahalad Thota (2020), Head of
Enterprise Analytics and Data Science at Wells Fargo, “We created an Artificial
Intelligence Program Bank which consisted of three different teams: the first was the
business team and their mandate was to identify the big use cases that we want to go
after and what are those big focus areas and figure out the areas they want to

5
 understand and see where they can apply AI and machine learning; the second team
was all about data and science and we ensure to bring the right data, identify the
problems and then make sure we have the right team members to be able to do the
model development; the third team was related to technology – we wanted to bring
these three groups together and drive forward the application of AI in the bank”.
ii. The Big Idea: Wells Fargo has applied machine learning and analytics in order to
identify and adjust to advanced fraud attacks in real-time. Matt Powell (2022), Head
of Fraud and Claims Management at Wells Fargo, mentions “As a company, we’re
focused on highly effective fraud prevention measures that adapt to constantly
changing threats with minimal disruption to the customer experience; our customers
expect and deserve banking services with strong protections from fraudsters that, at
the same time, are easy, convenient and accessible”. Wells Fargo has promised that
their machine learning model technology will adapt to future changes in fraud attacks.
In conclusion, data analytics is an important part of the banking industry, including Wells
Fargo, when it comes to meeting its financial goals and making sure that their customers are
receiving excellent services. Prahalad Thota (2020) stated that “to be able to do analytics at
any place, you need to have the right data and that data must be organized in a shape and
form that can be meaningfully used and explored to execute analytics”. Wells Fargo
continues to invest in technology and in its people, assuring that they are fully up to date
when it comes to combatting fraud.

Machine Learning
Machine learning models are in essence with artificial intelligence proficiencies, including
software that sanction intelligent engagement and process automation (Akanbi, Fazeldehkordi, &
Amiri, 2015) and it is known to be applied in cyber security to automate tedious tasks, encounter
cyber-attacks early on and disclose network weaknesses. It implements the ability to distinguish
significant patterns in data and has become a recurrent tool for almost any task faced with the
necessity of withdrawing meaningful information from data sets (Leo, Sharma & Maddulety,
2019). As Agus Sudjianto (no date), Head of Corporate Model Risk at Wells Fargo, said,
“Managing machine learning model risk is of the utmost importance in heavily regulated
industries such as finance; in particular, to manage potential risks due to bias/fairness, conceptual
soundness, implementation and model change control”.
i. Cyber Security: In banking, machine learning algorithms can inspect transactional
data in real time, recognizing dubious activities that contradict traditional patterns.
This proactive approach permits banks to discover possible threats before they occur
which can notably lessen the risk of financial loss. Cyberbanking and cyber security
are terms used to illustrate technologies, practices and processes that secure data,
networks and computer programs from cyber-attacks (Fréminville, 2020). Wells
Fargo has a webpage dedicated to giving tips on fraud prevention and cybersecurity

6
 for its customers including ways to protect their accounts, to protect themselves
online and to protect the elderly.
a. Data Integrity: Data integrity is the validity, entirety and status of data as it is
sustained over time and across layouts – guarding the integrity of the
organization’s data is a persistent operation. Wells Fargo tends to prioritize data
integrity in order to create a safe and reliable environment for their customers.
Encryption is a major form of data integrity within Wells Fargo, ensuring that the
data shared during online transactions is confidential and protected against any
fraudulent activity. Wells Fargo also takes into serious consideration data
validation, undergoing vigorous processes to verify the given data, even as far as
implementing automatic authentication checks.
b. Security Awareness Training: This training can assist organizations in minimizing
the risk of data breaches, malware ailments, phishing attempts and other hostile
schemes. By issuing employees with the apprehension and expertise needed to
stay safe and aware online, organizations can guarantee that their data is protected
and guarded from cyber-attacks while also reassuring their customers that they
can rely on the employees for being responsible. There are different ways Wells
Fargo takes into consideration security awareness training, two of them
incorporating guidelines for data protection and the collaboration with external
agencies such as FICO. Guidelines for data protection can help employees
become aware on how to safely handle their customer’s and organization’s
security data. Wells Fargo has a history of using FICO to fight against fraud to
protect their customers and consumers.

ii. Blockchain: Blockchains are allocated ledgers that can preserve secure valid records
and evidence of transactions, decreasing double the spending of tracked transactions.
It is a peer-to-peer network, dispensed registry system or database that sustains a list
of transactions rooted by attached nodes (mining nodes) without a third-party
mediator (Daluwathumullagamage & Sims, 2021). Blockchain technology plays a
role in Wells Fargo, enhancing its security to combat fraud. Wells Fargo has been
able to amplify traceability, processing all the transactions on blockchains to be
highly transparent making it difficult to miss any fraudulent activity.
a. Fraud Reduction: The potentiality of blockchain alleviating fraud is receiving a
lot of attention especially since 45% of financial arbitrators, such as stock
exchanges and money transfer services, undergo economic crime every year. A
majority of banking systems worldwide are constructed on a consolidated
database which puts them in a really endangering spot for cyberattacks because
instead of having many points of failure, they only have one point of failure; once
hackers reach that one breach, then they fully have access (Marr, 2017). This is
when the blockchain plays a role in preventing cyberattacks where each block
holds a timestamp and detains bundles of transactions with a link to a previous
block (Marr, 2017).

7
 b. Know Your Customer (KYC): Financial institutions are known to spend between
$60 million up to $500 million per year to stay up to date with Know Your
Customer (KYC) and customer assessments based on a Thomson Reuters Survey
(Marr, 2017). These proclamations are deliberate in decreasing money laundering
and terrorism schemes by utilizing directives for organizations to verify and
identify their clients. This leads to the depletion of administrative costs for assent
departments becoming notable.
In conclusion, machine learning has helped Wells Fargo come a long way in combating
fraudulent activities. They have gone to great lengths to take advantage of the advancement in
technology and create ways to protect their business and their customers. As mentioned on the
Wells Fargo (no date) website, “At Wells Fargo, the security of your assets is a top priority, and
we’re committed to protecting your company’s financial information and assets – we’re
proactively advancing our security to identify new threats and help ensure the safety of your
accounts and information”.

Predictive Analytics
Predictive analytics is defined as “an application of advanced statistical, information
software, or operations research methods to identify predictive variables and build predictive
models to identify trends and relationships not readily observed in the descriptive analytic
analysis” (Schniederjans, Schniederjans, & Starkey, 2014). When it comes to predictive
analytics, the Wells Fargo mobile app has an AI-powered predictive analytics tool. The bank
assures that this tool allows its customers to review an analysis of their spending habits and even
get recommendations on saving money based on their habits (Mejia, 2019).
Methods of Predictive Analysis:
i. Data Mining: Is defined as “a discovery-driven software application process that
provides insights onto business data by finding hidden patterns and relationships in
big or small data and inferring rules from them to predict future behaviors”
(Schniederjans, Schniederjans, & Starkey, 2014). These patterns and rules are known
to lead when it comes to decision-making. Wells Fargo uses data mining techniques
as part of their strategy when it comes to detecting and preventing fraud.
a. Transaction Monitoring: Data mining algorithms have been used to study
transaction data to identify unusual activity such as random spending locations,
attempts to overspend, and more. Wells Fargo uses these algorithms to
consistently monitor customer transactions such as purchases, withdrawals and
wire transfers in order to look out for unordinary patterns. They have 24/7 fraud
monitoring and once they detect any fraud, they immediately contact the
customer. Some of the steps they go through are sending a verification code to
make sure that the customer is the one making the purchase, declining
transactions and limiting online banking (Wells Fargo, no date).

8
 b. Account Profiling: Data mining algorithms can help in differentiating customer
transactions against the customer profiles that show any unnecessary behavior
compared to their usual behavior in a purchase. Wells Fargo creates profiles for
each of their customers regarding their financial behavior, demographics and
spending history which helps them identify any fraudulent activity that goes
against their account. When creating an account with Wells Fargo, it is strongly
advised to use a unique username and password, permitting customers to have a
strong security force for their account. They also use a tool known as Advanced
Access which they describe as “an additional layer of security that helps protect
your information and prevent unauthorized transactions” (Wells Fargo, no date).

ii. Regression Analysis: According to Catherine Cote (2021), regression analysis is “the
statistical method used to determine the structure of a relationship between two
variables (single linear regression) or three or more variables (multiple regression)
and it is used for two primary purposes: (1) to study the magnitude and structure of
the relationship between variables and (2) to forecast a variable based on its
relationship with another variable”. Wells Fargo uses regression analysis to identify
fraud detection by looking into the customer’s data that is vulnerable to fraud.
a. Risk Scoring: Wells Fargo assigns risk scores to certain transactions or customer
profiles. If a transaction shows any potential elements of fraudulent activity, then
Wells Fargo uses this model to assign a high-risk score to that transaction (Wells
Fargo, no date). By doing this, it categorizes these alerts for further investigation
and puts them into perspective for future transactions.
b. Variable Selection: Regression analysis lends a hand in selecting specific
variables that help predict fraud. Once Wells Fargo has become aware of which
variables are very prominent, it can then focus its attention in observing and
investigating those specific features in customer behavior and transactions (Wells
Fargo, no date).
Wells Fargo turned to predictive analytics in order to strengthen their abilities in
combating fraud and upgrading their security systems. They have even been able to include
predictive analytics applications in smartphones, allowing customers to have access to that tool
when it comes to their personal banking. By doing so, Wells Fargo has proceeded to give
customers the best experience through their phones, giving them the ability to provide a more
eligible and accessible way of helping their consumers.

9
 Part 2: Credit Card Approvals
Introduction
The process of getting approved for a credit card is usually pretty straightforward. The
provider runs a credit check on the finances before looking into approving the application along
with the interest rate that will be charged and the credit limit given. Debt is defined as money
that is owed or due and it is no secret that many people, unfortunately, undergo debt considering
their circumstances. The tables and graphs in this section will go over credit cards approvals,
specifically how debt is affected depending on the years of employment. The dataset used was
taken from Kaggle (https://www.kaggle.com/datasets/samuelcortinhas/credit-card-approval-
clean-data/data) which contains important variables such as age, years of employment, debt,
income, credit score and more.
Regression analysis is “a statistical tool that allows us to describe the way in which one
variable is related to another and it may be a simple one involving just two variables in a single
equation or it may be very complex having many variables and equations” (Wilson, Keating, &
Beal-Hodges, 2012). Regression tables were used within these hypotheses, in which they are
meant to go against the null hypothesis and prove it wrong. Along with the regression tables,
scatter plots have been included to display the equation and R Square value using a simple linear
regression which is a model that involves a single independent variable.
The descriptive statistics analysis will give details of the selected variables and create a
summary between the two. A figure of a table will be included showing the different patterns in
the Y variable (dependent) and the X variable (independent). The dependent variable, also
known as the response variable, is the variable that researchers try to explain or predict because
it depends on another variable. The independent variable, also known as the explanatory
variable, is the variable researchers use to explain or predict the other variable’s (dependent)
value.

10
Descriptive Statistics Analysis of Selected Variables
Descriptive statistics refers to “the analysis, summary and presentation of findings related
to a data set derived from a sample of entire population” (CFI Team, 2023). These statistics
consists of three main sections – central tendency, frequency distribution and variability. Figure
1 below analyzes the years of employment and debt describing their mean, median, mode,
standard deviation, sample variance, kurtosis, skewness, range, and confidence level.
Years Employed: The average
number of years employed is 2.2 years
with a standard deviation of 3.3 years,
indicating a considerable variation. The
median years employed is 1 year with
the highest number of years employed
being 28.5 years. The skewness level is
at 2.89 while the kurtosis level is at
11.20, showing that the distribution of
years employed is positively skewed
with longer tails.
Debt: The average number of
debts is £4,758 with a standard
deviation of £4,978, indicating a
considerable variation. The median debt
is £2,750 with the highest number of
debts being £28,000. The level of
skewness is 1.48 with a kurtosis level of
2.27, showing that the distribution of
Figure 1 Descriptive Analysis of Selected Variables
debt is positively skewed with longer tails.
In conclusion, the descriptive statistics show an extensive rundown of debt within years
of employment. This information is very useful for banks when it comes to credit card approvals
and looking into how much debt someone has depending on their years of employment. Banks
could then look into ways on how they can assist those in debt and improve their chances of
getting a credit card approval.

11
Alternate hypothesis: Years of employment do not have a huge effect on debt when
it comes to credit card approvals and non-approvals – will be split into two parts
Hypothesis A: Within approved and non-approved credit card applications, debt did not
have much of an effect depending on the years of employment ranging from 0 to 28 years.
Below will contain a regression table (Figure 2) and a scatter plot (Figure 3) for this
hypothesis.

Figure 2 Regression Table: Hypothesis A

Within this dataset, the Y variable 30

determines the debt and the X variable
determines the years of employment. Given the 25
regression statistics, the R Square is at 8%, only
explaining 8% of the variance observations (the 20
y = 0.4446x + 3.7701
behavior) which is not a big percentage. The R R² = 0.0893
15
Square is known as the “goodness-of-fit”
measure “indicating the percentage of the
10
variance in the dependent variable that the
independent variable explains collectively” 5
(Frost, 2023).
0
For the graph, a simple linear regression
0 5 10 15 20 25 30
was created to showcase the equation
y = 0.4446x + 3.7701. Figure 3 Scatter Plot Graph: Hypothesis A

The correlation between debt and years of employment equaled 0.298902. Considering
the P Value is small, that means the model is statistically significant.
The Significance F, in number terms, should be less than 0.05 which goes against the null
hypothesis meaning that it is not caused by random events. In this regression table, the
Significance F is at 0.00. The Standard Error is a measurement of the unpredictability of the
model – if it is right or wrong – and the ideal value to have within this measurement is around 0

12
or 1. As shown on the table, the Standard Error is low on this table (0.2), meaning the given
coefficient estimate is accurate (0.44).
When the X variable increases by 1 unit, the Y variable will increase by 0.44 which is the
coefficient. Since the coefficient is positive, then the X variable increases as well as the Y
variable.
According to the predictive analysis, it was found that the trendline is expressed by y=
0.4446x + 3.7701. When the years of employment equal to 0, the debt equals to £3,770.
This concludes that the R Square in this circumstance is not very strong, considering that
the percentage is at 8% which is not a good fit and is a weak correlation.

13
 Hypothesis B: Within approved and non-approved credit card applications, debt did have
a slightly higher effect depending on the years of employment ranging from greater than or equal
to 1 year, 28 years being the most.
Below will contain a regression table (Figure 4) and a scatter plot (Figure 5) for this
hypothesis.

Figure 4 Regression Table: Hypothesis B

Just as in the previous dataset, the Y 30

variable in this dataset still determines the debt
and the X variable also determines the years of 25
employment. As shown in this regression table,
the R Square is a bit higher than the previous 20
y = 0.4552x + 3.7102
hypothesis, displaying at 11%. R² = 0.1194
15
For the graph, a simple linear regression
was created to showcase the equation 10
y = 0.4552x + 3.7102.
5
The correlation between debt and years
of employment equaled 0.345542, this 0
correlation also being at a higher number than 0 5 10 15 20 25 30
the first. Since the P Value is low then the Figure 5 Scatter Plot Graph: Hypothesis B
model is statistically significant.
The Significance F in this table is at 0.00 and since it is below 0.05, that means that this
goes against the null hypothesis which is not caused by random events. The Standard Error is
low on this table, equaling to 0.3, meaning that the given coefficient is accurate (0.45).
When the X variable increases by 1 unit, then the Y variable will increase by 0.45 (the
coefficient). Considering the coefficient is positive, when the X variable increases, the Y variable
also increases.

14
 According to the predictive analysis, it was found that the trendline is expressed by y=
0.4552x + 3.7102. When the years of employment equal to 0, the debt equals to £3,710.
This concludes that the R Square in this circumstance is a little higher but not very
strong, showing that the percentage rose to 11% which is still not a good fit and is considerably a
weak correlation.
In the overall conclusion, removing the length of employment that had not reached a year
yet improved the model by 3%. The comparison within these regressions is around the same and
similar, not displaying a very big jump.

15
Recommendations for Enhancing Analytical Skills to Better Combat Fraud
Within Wells Fargo
With Wells Fargo already being far ahead in technology when it comes to fraud prevention,
there are some ways they can introduce to their teams to improve and better their abilities:
• Periodic Skills Assessments: Undergoing periodic skills assessments can help the team in
staying up to date with their skills and assess their actions in where they can improve. By
doing so, they will be well informed in the new ways that fraudulent malice can interrupt
their company.

• Use Augmented Reality for Practice Case Studies: Utilizing AR to create scenarios
within simulations for team members to practice case studies can go a long way in
strengthening their awareness when it comes to fraudulent activities. Putting themselves
in that situation makes them analyze what actions they can take in order to prevent fraud
within their company and customers.

• Advanced Training Programs: In providing these programs, employees can dive deep into
the fraudulent world and learn more about advanced technology, such as ML and AI, and
discover better ways to apply these technologies. This training may even open doors in
figuring out how to better that technology and come out with upgraded versions.

16
Recommendations for Improving the Analysis Within the Dataset
Even though the dataset did include a good display of variables to use in analyzing the
information, some recommendations can be made in order to further look into this data:
• Include More Variables: The dataset includes basic variables that are useful for the
overall necessary details, but it would be good to incorporate more specific variables in
order to get more personalized outcomes. These variables can be area of living (suburbs,
city, etc.), title at work (VP, CEO, Admin, etc.) and number of children – all of these
variables clearly having an effect on debt.

• Allow Room to Compare Industries: With the industries already being a part of the
dataset, it would have been more interesting to have the ability to compare the industries
and how much debt is collected within those industries. This could have played a major
role in deciphering who is more vulnerable to falling victim to debt.

• Specify Which Banks: When specifying which banks people use, it could specifically
display the banks that are more susceptible to debt. Exploring this on a deeper level can
lead to investigating credit cards and their sensitivity depending on debt.

17
Conclusion
In the first part, this report has shed light on business analytics and its importance in
combatting fraud specifically within banks, such as Wells Fargo. Many forms of business
analytics were researched and explained, from definitions to ways Wells Fargo incorporates such
tactics. Data Analytics, Machine Learning and Predictive Analytics have all been analyzed,
proving that these advancements in technology play a major key role in the financial industry
and fending off fraudulent activities.
In the second part, a dataset of credit card approvals was studied specifically looking into
years of employment and debt. As the analysis was looked over, predictions came about leading
to investigating different hypotheses. With the findings, it puts into perspective what can be done
when it comes to approving credit cards, especially if specific banks play a role in approving
them no matter the debt.
Recommendations were constructed to better enhance analysis within the banks fighting
off fraud and the approvals of credit cards overlooking debt. By including certain tactics within
the banks, they would have a better idea in preventing fraud by expanding their abilities. As for
the graphs, developing more specific variables can help identify where debt can really be
affected. In doing all of this, a new world would be opened up with analysis and this industry can
undergo a positive development.
It has been shown that by implementing the use of data analytics and continuing to
discover new ways on how to improve and advance, banks can go a long way in keeping their
systems and customers safely guarded. Throughout the years, banks have examined various
processes on how to protect themselves internally and externally, leading them to the number of
technological creations that have deemed them sophisticated in that matter.

18
Bibliography
Subramanian, R 2014, Bank Fraud: Using Technology to Combat Losses, John Wiley & Sons,
Incorporated, Newark. Available from: ProQuest Ebook Central. [1 December 2023].
Choudhury, M.D. and Choudhury, M.D. (2020) 'Why data and analytics is so significant for
Wells Fargo,' Express Computer, 6 January. https://www.expresscomputer.in/news/why-data-
and-analytics-is-so-significant-for-wells-fargo/44940/.
Schniederjans, M.J., Schniederjans, D.G. and Starkey, C.M. (2014) Business Analytics
Principles, Concepts, and Applications: What, Why, and How. Pearson Education.
Daluwathumullagamage, Dulani Jayasuriya; Sims, Alexandra (2021): Fantastic beasts:
Blockchain based banking, Journal of Risk and Financial Management, ISSN 1911-8074, MDPI,
Basel, Vol. 14, Iss. 4, pp. 1-43, https://doi.org/10.3390/jrfm14040170.
Marr, B. (2017) 'Practical examples of how blockchains are used in banking and the financial
services sector,' Forbes, 10 August.
https://www.forbes.com/sites/bernardmarr/2017/08/10/practical-examples-of-how-blockchains-
are-used-in-banking-and-the-financial-services-sector/?sh=181dd3ab1a11.
Protecting Customers (no date). https://www.aba.com/banking-
topics/technology/cybersecurity/protecting-customers.
Protecting you and your accounts | Wells Fargo (no date). https://www.wellsfargo.com/privacy-
security/fraud/protecting-you/.
Mejia, N. (2019) Artificial intelligence at Wells Fargo – A brief overview. https://emerj.com/ai-
sector-overviews/artificial-intelligence-at-wells-fargo/#:~:text=as%20account%20balances%3A-
,Predictive%20Analytics%20Tool%20for%20Mobile%20Banking,money%20based%20on%20t
hose%20habits.
What is regression analysis in business analytics? (2021).
https://online.hbs.edu/blog/post/whatis-regression-analysis.
Leo, M., Sharma, S. and Maddulety, K. (2019) 'Machine Learning in Banking Risk
Management: A Literature Review,' Risks, 7(1), p. 29. https://doi.org/10.3390/risks7010029.
Annual LexisNexis Risk Solutions Study Finds Fraud Costs up to 22.4% from Pre-Pandemic
Levels Across U.S. and Canadian Financial Services Firms | LexisNexis Risk Solutions (no
date). https://risk.lexisnexis.com/about-us/press-room/press-release/20221116-study-finds-fraud-
costs#:~:text=Costs%20continued%20to%20rise%20above,every%20%241%20lost%20to%20fr
aud.
Gupta, T. et al. (2019) 'Role of Big Data Analytics in Banking,' Role of Big Data Analytics in
Banking [Preprint]. https://doi.org/10.1109/ic3i46837.2019.9055616.
Adrian, B. (2015) 'Detecting and Preventing Fraud with Data Analytics,' Procedia. Economics
and Finance, 32, pp. 1827–1836. https://doi.org/10.1016/s2212-5671(15)01485-9.

19
Team, C. (2023) Descriptive statistics. https://corporatefinanceinstitute.com/resources/data-
science/descriptive-statistics/.
Wilson, J.. H., Keating, B. and Beal-Hodges, M. (2012) Regression Analysis: Understanding and
building business and economic models using Excel. https://doi.org/10.4128/9781606494356.
Frost, J. (2023) How to interpret R-squared in regression Analysis.
https://statisticsbyjim.com/regression/interpret-r-squared-regression/.
Wells Fargo enhances fraud protection and customer experience using FICO solutions (2022).
https://www.nasdaq.com/press-release/wells-fargo-enhances-fraud-protection-and-customer-
experience-using-fico-solutions.
Wells Fargo: Interpretable Machine Learning | H2O.ai (no date). https://h2o.ai/case-
studies/interpretable-machine-learning/.
Fréminville, M. (2020) Cybersecurity and decision makers.
https://doi.org/10.1002/9781119720362.
Akanbi, O.A., Fazeldehkordi, E. and Amiri, I.S. (2015) A Machine-Learning Approach to
Phishing Detection and Defense, Elsevier eBooks. https://doi.org/10.1016/c2014-0-03762-8.
Javaid, M. et al. (2022) 'A review of Blockchain Technology applications for financial services,'
Bench Council Transactions on Benchmarks, Standards and Evaluations, 2(3), p. 100073.
https://doi.org/10.1016/j.tbench.2022.100073.
Lin, N. (2014) Applied Business Analytics: Integrating Business Process, Big Data, and
Advanced Analytics. FT Press.

20