Project Proposal

Rational

Credit card fraud detection is presently the most frequently occurring problem in the present
world which results in huge financial losses. The number of online transactions has grown in
larger quantities & online credit card transactions hold a huge share of these transactions.
Therefore, banks & financial institutions offer credit card fraud detection applications much
value & demand. Credit card fraud generally happens when the card was stolen for any of the
unauthorized purposes or even when the fraudster uses the credit card information for his use.

It is vital that credit card companies are able to identify fraudulent credit card transactions so that
customers are not charged for items that they did not purchase. Such problems can be tackled
with Data science and its importance, along with Machine learning, cannot be overstated. This
project intends to illustrate the modeling of a data set using machine learning with Credit Card
Fraud Detection. This problem includes modeling past credit card transactions with the data of
the ones that turned out to be fraud. This model is used to recognize whether a new transaction is
fraudulent or not. Our objective here is to detect 100% of the fraudulent transactions while
minimizing the incorrect fraud classifications. In this process, we have focused on analyzing and
pre-processing data sets as well as the deployment of multiple anomaly detection algorithms
such as Local Outliner Factor and Isolation Forest algorithm on the PCA transformed Credit
Card Transaction data.

Introduction

Credit card fraud is a growing concern with far reaching consequences in the government,
corporate organizations, finance industry, In Today’s world high dependency on internet
technology has enjoyed increased credit card transactions but credit card fraud had also
accelerated as online and offline transaction. As credit card transactions become a widespread
mode of payment, focus has been given to recent computational methodologies to handle the
credit card fraud problem. There are many fraud detection solutions and software which prevent
frauds in businesses such as credit card, retail, e-commerce, insurance, and industries. Machine
Learning is one notable and popular methods used in solving credit fraud detection problem. It is
impossible to be sheer certain about the true intention and rightfulness behind an application or
transaction. In reality, to seek out possible evidences of fraud from the available data using
mathematical algorithms is the best effective option. Fraud detection in credit card is the truly the
process of identifying those transactions that are fraudulent into two classes of legit class and
fraud class transactions, several techniques are designed and implemented to solve to credit card
fraud detection such as genetic algorithm, artificial neural network frequent item set mining,
migrating birds optimization algorithm, comparative analysis of decision tree and random forest
is carried out. Credit card fraud detection is a very popular but also a difficult problem to solve.
Firstly, due to issue of having only a limited amount of data, credit card makes it challenging to
match a pattern for dataset. Secondly, there can be many entries in dataset with truncations of
fraudsters which also will fit a pattern of legitimate behavior. Also the problem has many
constraints. Firstly, data sets are not easily accessible for public and the results of researches are
often hidden and censored, making the results inaccessible and due to this it is challenging to
1
benchmarking for the models built. Datasets in previous researches with real data in the literature
is nowhere mentioned. Secondly, the improvement of methods is more difficult by the fact that
the security concern imposes a limitation to exchange of ideas and methods in fraud detection,
and especially in credit card fraud detection. Lastly, the data sets are continuously evolving and
changing making the profiles of normal and fraudulent behaviors always different that is the legit
transaction in the past may be a fraud in present or vice versa. This paper evaluates two advanced
machine learning, Decision tree and random forests and then a collative comparison is made to
evaluate that which model performed best. Credit card transaction datasets are rarely available,
highly imbalanced and skewed. Optimal feature (variables) selection for the models, suitable
metric is most important part of mining to evaluate performance of techniques on skewed credit
card fraud data. A number of challenges are associated with credit card detection, namely
fraudulent behavior profile is dynamic, that is fraudulent transactions tend to look like legitimate
ones, Credit card fraud detection performance is greatly affected by type of sampling approach
used, selection of variables and detection technique used.

Literature Survey

In this section, we briefly review the related work on credit card fraud system and their different
techniques. In this document proposes a new comparative measure of the comparison rules that
reasonably represents the profits and losses due to fraud detection. A cost-sensitive method based
on the minimum Bays risk is presented using the proposed cost measure. Improvements of up to
23% are obtained by comparing this method and other latest-generation algorithms. The data set
for this document is based on the real-life transactional data of a large European company and
personal data in the data is kept confidential. The accuracy of an algorithm is about 50%. The
importance of this work was to find an algorithm and reduce the cost measurement. The result
was 23% and the algorithm they found was the minimal risk of Bays. In Several modern
techniques based on sequence alignment, machine learning, artificial intelligence, genetic
programming, data mining, etc. They have been developed and are still being developed to detect
fraudulent credit card transactions. A solid and clear understanding of all these approaches is
needed, which will undoubtedly lead to an efficient credit card fraud detection system. This
document shows a survey of different techniques used in credit card fraud detection mechanisms
and the evaluation of each methodology based on certain design criteria. An analysis of credit
card fraud detection methods was performed. The survey in this document was based solely on
detecting the efficiency and transparency of each method. The importance of this document was
to conduct a survey to compare different credit card fraud detection algorithms to find the most
appropriate algorithm to solve the problem. In A comparison was made between models based
on artificial intelligence together with a general description of the fraud detection system
developed in this document, such as the naive Bayesian classifier and the Bayesian network
model, the clustering model. And finally, conclusions are drawn on the results of the model
evaluation tests. The number of legal truncation was determined to be greater than or equal to
0.65, ie its accuracy was 65% using the Bayesian network. The importance of this document is to
compare the models based on artificial intelligence together with a general description of the
developed system and to establish the accuracy of each model together with the recommendation
to create the best model. In Nutan and Suman on review on credit card fraud detection they have
supported the theory of what is credit card fraud, types of fraud like telecommunication,
bankruptcy fraud etc. and how to detect it, in addition to it they have explained numerous
algorithms and methods on how to detect fraud using Glass Algorithm, Bayesian, networks,
2
Hidden Markova model, Decision Tree and 4 more. They have explained in detail about each
algorithm and how this algorithm works along with mathematical explanation. Types of machine
learning along with classifications have been studied. Pros and cons of each method is listed

Problem Statement

To build credit card fraud detection system using machine learning algorithms. The major aim of
this project is to perform a comprehensive review of different fraud detection methods & some
innovative machine learning techniques.

Proposed Methodology

In this system evidences from current as well as past Behavior are combined. A fraud detection
system is proposed that includes rule based filter, Dempster Shafer adder, transaction history
database and Bayesian learner. In rule base the suspicion level of each incoming transaction is
determined. Dumpster Shafer is used to combine multiple such evidences and an initial belief is
computed. Based on this belief the transactions are classified as normal, abnormal or suspicious.
The incoming transactions are initially handled by the rule base using probability values. After
this the values are combined using Dumpster Shafer Adder. If the transaction is declared as
fraudulent then it is handled by the card holder. If suppose the transaction is suspicious then it is
fed in the suspicious table. The score of transaction is updated in the database with the help of
machine learning classification. This architecture is flexible such that new kinds of fraud can be
handled easily. With the help of learner the system can dynamically adapt to the changing needs.

3
Resource Requirement:

Hardware Requirements:

Sr. No Resources Configuration

1 Processor Intel® Pentium® CPU B960
2. Speed 2.20 GHz
3. RAM 4 GB
4. Key Board Standard Windows Keyboard
5. Mouse Two or Three Button Mouse

Software Requirements:

Sr. No. Resources Configuration

1. Operating System Windows 8.1 pro
2. Coding Language Python
3. Software Python

Active Plan / Project Plan:

Sr no. Duration Tasks

1. 15-September-2021 To 25-September-2021 Selection of domain
2. 29-September-2021 To 06-October-2021 Topic selection
3. 07- October -2021 To 23- October -2021 Base paper selection
4. 25- October -2021 To 30- October -2021 Existing system
5. 03-November-2021 To 17- November -2021 Requirement collection from
Industry
6. 25- November -2021 To 15-December-2021 Project proposal plan.
7. 1st week of January Presentation on project Plan
8. 14-February-2022 To 21- February -2022 Module design.
9. 22- February -2022 To 01-March-2022 GUI.
10. 02- March -2022 To 30- March -2022 Implementation
11. 01-April-2022 To 11-April-2022 Testing.
12. 15-April-2022 To 30-April-2022 Installation of application
13. Last week of may Final project presentation