Subject: FAM (22593) Academic Year: 2023 -24

Branch Name: AN-5-I Semester: Fifth

Credit Card Fraud Detection Using Machine Learning

By the group of 4 students

Roll No Name Of Students Enrollment Seat No

(Sem-5) No.
30 Mr. Yash Patil 2216290253 149167

23 Mr. Sameer Jadhav 2116290142 149161

11 Mr. Aditya Kesari 2116290128 149149

21 Mr. Masir Momin 2116290140 149159

Under the guidance of

Ms. Renuka Dingore
AT
Department of Artificial Intelligence and Machine Learning
(Diploma)

MAHARASHTRA STATE BOARD OF TECHNICAL EDUCATION

 CERTIFICATE

This Micro Project Report Entitled “Credit Card Fraud Detection Using Machine
Learning” By Mr. Yash Patil, Mr. Sameer Jadhav, Mr. Aditya Kesari, Mr.
Masir Momin Is Approved for the Diploma of AIML Engineering (Diploma)
(Third Year) For Academic Year 2023 - 2024.

Examiners

1.

2.

Supervisor

1.

Ms. Renuka Dingore

Head of the Department Principal

Date:

Place:
 ACKNOWLEDGEMENT

A micro project is something that could not have been materialized

without cooperation of many people. This project shall be incomplete if I
do not convey my heartfelt gratitude to those people from whom I have
got considerable support and encouragement. It is a matter of great
pleasure for us to have respected Ms.Renuka Dingore as my project guide.
We are thankful to her for being a constant source of inspiration. We
would like to give our sincere thanks to Prof. Heena Patil Head of
Department, for their kind support.
 CONTENTS

Sr. No Name Of Topic Page

No.
1 Abstract 5

2 Introduction 6

3 Aim of Project 7

4 Types of credit card fraud 8-9

5 Credit card fraud detection techniques 10-11

6 Advantages & Disadvantages 12-13

7 Implementation 14

8 Result 15

9 Conclusion 16

10 References 17
 ABSTRACT

It is a systematic literature review to reflect the previous studies that dealt with
credit card fraud detection and highlight the different machine learning techniques to
deal with this problem. Credit cards are now widely utilized daily. The globe has just
begun to shift toward financial inclusion, with marginalized people being introduced
to the financial sector. As a result of the high volume of e-commerce, there has been
a significant increase in credit card fraud. One of the most important parts of today's
banking sector is fraud detection. Fraud is one of the most serious concerns in terms
of monetary losses, not just for financial institutions but also for individuals. as
technology and usage patterns evolve, making credit card fraud detection a
particularly difficult task. Traditional statistical approaches for identifying credit card
fraud take much more time, and the result accuracy cannot be guaranteed. Machine
learning algorithms have been widely employed in the detection of credit card fraud.
The main goal of this review intends to present the previous research studies
accomplished on Credit Card Fraud Detection (CCFD), and how they dealt with this
problem by using different machine learning technique
 INTRODUCTION

The payment industry is increasingly offering digital payment methods for a variety
of reasons, including time savings, the ability to pay for purchases over time, the
expansion of the market, ease of use, convenience, credit card rewards, price
protection, purchase protection, and travel benefits, However, it is susceptible to
internet fraud and may increase business expenditures. unfortunately, With the
current trend of financial inclusion, increasingly offering digital payment methods,
and marginalized people being introduced to the financial sector, the number of card
users increases, so does the revenue, making them more vulnerable to fraud
 AIM OF MICRO-PROJECT

The aim of this project is to develop an effective and efficient credit card fraud
detection system utilizing machine learning techniques. By harnessing the power of
advanced data analytics and predictive modeling, this project seeks to mitigate
financial losses for both credit card users and financial institutions, enhance security
in the digital payment ecosystem, and contribute to the ongoing battle against
fraudulent activities. The primary objective is to build a robust and accurate fraud
detection model capable of identifying suspicious transactions in real-time, thereby
safeguarding the financial interests and data privacy of consumers while minimizing
false positives, and ultimately providing a secure and seamless experience for
legitimate
 TYPES OF CREDIT CARD FRAUD

Types of credit card fraud

1.Behavioural Fraud: The first type of credit card fraud is a Behavioural Fraud that
has four types introduced:

1.1 Stolen/Lost Card: Criminals steal a credit card to get access to a lost card
1.2 Mail Theft: Before reaching the real user, fraudsters receive credit cards in the
mail.
1.3 Counterfeit: Fraudsters take card information from one source and use it on
websites that do not require a physical card.
1.4 Merchant Collusion and Triangulation: Criminals pose as an intermediary site,
collecting customer credit card information and redirecting orders using stolen credit
card information.

2.Application Frauds: The second type of credit card fraud is application fraud which
has six types introduced:
2.1 Account Theft and Suspicious Transactions: Criminals can utilize personal
information such as a Social Security number, a secret question answer, or a date of
birth taken from an individual to conduct financial transactions. Because identity theft
is linked to a large number of fraud transactions, financial fraud detection systems
should focus on establishing an analysis of a user's behavior.
2.2 Clone transactions: Among the various kinds of credit card frauds, clone
transactions are common. It simply refers to replicating or making transactions that
are similar to the original. When an organization tries to collect payment from a
partner many times by sending the same invoice to different departments, this can
happen.
2.3 Credit Card Skimming (electronic or manual): Making an illegitimate clone of a
credit or bank card with a device that reads and copies information from the original
card is known as credit card skimming or credit card forging. Scammers extract card
numbers and other credit card information with equipment known as "skimmers,"
save it and resale it to criminals.
2.4 CNP (Card Not Present) Fraud: This could happen if criminals discover the card's
expiration date and account number. These two factors are critical when making an
internet purchase. More retailers are requiring the verification code these days, but it
is not difficult to obtain if you know your account number and expiration date.
Criminals can simply try to enter the verification code at a low frequency and
eventually figure it out. Anomaly detection tools, such as Machine Learning, may be
useful in detecting suspicious trends in a client's activity in order to combat this form
of fraud.
2.5 Phishing: it is a very common type of data theft method. The victim receives a
legitimate looking email posing as a representative of a well-known organization. It
could be a request to update account information or transmit additional personal data
in response to "changes" in the organization's policy or for any other cause. The
victim transmits their personal information without paying enough
 CREDIT CARD FRAUD DETECTION TECHNIQUES

1.Decision tree : A supervised learning methodology, graphical representations of

possible solution choices based on certain situation’s As in Figure and it is a tree-
structured classifier. It starts with a roots node where inside nodes represent the
features of a datasets ,branches symbolize the decision instructions and each leaf node
represents the results. In decision tree and they have the purpose of deciding and
communicating respectively. A decision tree plainly asks a questions and then divide
it into sub trees based on the answers. Although DT can solve classifications and
regression problems. it is most commonly used to solve classification problems. To
find the datasets classes, the algorithm searches at the top of the tree. It compares the
root Trait with the record attribute and follow the offshoot on way to the next node,
which it calculate depending on the relations.
2.Random Forest : Random Forest classifiers finds decision trees in a subsets of the
data and then aggregates their details to that to get the full datasets predictive powers.
Rather than relying on a single decision tree. The RF take the predictions from each
tree and forecasts the final outputs based on the majority votes of forecasts. Using a
huge number of trees in the forest improves precisions and eliminates the issues of
over fitting. It predicts output with high precisions, and it runs efficiently even with
large datasets. It can also keeps accuracy when large proportions of data is lost.
Random Forest can handle both classification and regression task. It can handle large
datasets with high dimensionality
 ADVANTAGES AND DISADVANTAGES

 ADVANTAGES :

 Improved Accuracy: Machine learning algorithms can analyze vast amounts

of transaction data to identify subtle patterns and anomalies, leading to more
accurate fraud detection compared to rule-based systems.

 Real-time Detection: Machine learning models can operate in real-time,

enabling the immediate identification of fraudulent transactions and allowing
for prompt action to minimize potential losses.

 Adaptability: Machine learning models can adapt and evolve over time to
stay ahead of changing fraud tactics and patterns, making them more resilient
to new threats.

 Reduced False Positives: By analyzing various factors and considering the

context of transactions, machine learning can reduce the number of false
positives, resulting in a better user experience and reduced operational costs
for investigating false alarms.
 DISADVANTAGES :

 Overfitting: Machine learning models can overfit to the training data, which
may result in high accuracy on the training set but poor generalization to new,
unseen data.

 Costly Implementation: Developing, deploying, and maintaining machine

learning systems can be costly, particularly for smaller financial institutions
with limited resources.

 Model Complexity: Some machine learning models can be complex and

difficult to interpret, making it challenging to explain and justify the reasons
for a particular fraud detection decision.

 Model Updates: Models require continuous updates to adapt to evolving

fraud patterns, which can be time-consuming and resource-int
 IMPLEMENTATION

This idea is difficult to implement in real life because it requires the cooperation from
banks, which aren’t willing to share information due to their market competition,
and also due to legal reasons and protection of data of their users. Therefore, we
looked up some reference papers which followed similar approaches and gathered
results. As stated in one of these reference papers: “This technique was applied to a
full application data set supplied by a German bank in 2006. For banking
confidentiality reasons, only a summary of the results obtained is presented below.
After applying this technique, the level 1 list encompasses a few cases but with a high
probability of being fraudsters. All individuals mentioned in this list had their cards
closed to avoid any risk due to their high-risk profile. The condition is more complex
for the other list. The level 2 list is still restricted adequately to be checked on a case
by case basis. Credit and collection officers considered that half of the cases in this
list could be considered as suspicious fraudulent behaviour. For the last list and the
largest, the work is equitably heavy. Less than a third of them are suspicious. In order
to maximize the time efficiency and the overhead charges, a possibility is to include
a new element in the query; this element can be the five first digits of the phone
numbers, the email address, and the password, for instance, those new queries can be
applied to the level 2 list and level 3 list”.
 RESULT

The code prints out the number of false positives it detected and compares it with the
actual values. This is used to calculate the accuracy score and precision of the
algorithms. The fraction of data we used for faster testing is 10% of the entire dataset.
The complete dataset is also used at the end and both the results are printed. These
results along with the classification report for each algorithm is given in the output as
follows, where class 0 means the transaction was determined to be valid and 1 means
it was determined as a fraud transaction. This result matched against the class values
to check for false positives. Results with the complete dataset is used:
 CONCLUSION

Because of the seamless, simple, and convenient usage of e-commerce, digitalization

is growing popularity these days. It became a very common and simple form of
payment. People prefer e-payments and e-shopping. Since it is more convenient in
terms of time, transportation, and so on as a result of the massive use of e-commerce,
there has also been a significant increase in credit card fraud. Fraudsters try to Misuse
of a credit card and the lack of transparency in online payments are two issues that
need to be addressed. As a result, combating fraudsters' activities has become
extremely difficult. The major goal is to keep credit cards transactions safe so that
consumers can safely and easily utilize e-banking. In today's world, one of the most
important aspects of banking is fraud detection. Fraud is one of the most significant
concerns in terms of monetary losses, not only for merchants but also for individual.
Applying data quality dimensions and experimenting with different strategies to
partition datasets with machine learning algorithms will result in the best performance
and accuracy in detecting credit card fraud. to accurately identify frauds, quickly
discover fraud cases streaming, and the ability to detect online fraud in real-time,
requiring less time for variation approaches, and detecting hidden correlations in data.
The results of Traditional methods have no high accuracy, its results not guaranteed,
and it takes more time to process. Deep learning has recently been used to solve
complex issues in a variety of fields. The performance of deep learning methods for
credit card fraud detection is compared to machine learning algorithms, results reveal
that the deep learning methods outperform traditional machine learning models,
implying that the deep learning approaches can be used to detect credit card fraud in
real-world situations.
 REFERENCES

[1] S. H. Projects and W. Lovo, ―JMU Scholarly Commons Detecting credit

card fraud : An analysis of fraud detection techniques,‖ 2020.

[2] C. Reviews, ―a Comparative Study : Credit Card Fraud,‖ vol. 7, no. 19, pp.
998–1011, 2020

[3] A. Agarwal, ―Survey of Various Techniques used for Credit Card Fraud
Detection,‖ Int. J. Res. Appl. Sci. Eng. Technol., vol. 8, no. 7, pp. 1642–1646,
2020, doi:10.22214/ijraset.2020.30614.

[4] ―Machine Learning Decision Tree Classification Algorithm-Javatpoint.‖

https://www.javatpoint.com/machinelearning-decision-tree-classification-
algorithm (accessedApr. 03, 2021)[31] ―Machine Learning RandomForest
Algorithm Javatpoint.‖ https://www.javatpoint.com/machine-learning random-
forest-algorithm (accessed Apr. 03, 2021)

[5] B. Wickramanayake, D. K. Geeganage, C. Ouyang, and Y.Xu, ―A survey

of online card payment fraud detectionusing data mining-based methods,‖
arXiv, 2020