VIDYAVARDHAKA COLLEGE OF ENGINEERING

(Autonomous, affiliated to VTU)

DEPARTMENT OF

CSE (ARTIFICIAL INTELLIGENCE & MACHINE LEARNING)

Mini Project
Report On

“Credit Card Fraud Detection”

Submitted in partial fulfillment of the requirement for the completion of Activity Based
Learning (ABA) of subject Automata Theory [21CI/AI54] of V semester of

BACHELOR OF ENGINEERING

Submitted by:

ARATHI K 4VV20CI009
DEEKSHITA N 4VV20CI014
NAGAPRAPTHI P S 4VV21CI031
RESHMA R 4VV20CI042
SHREYA B S 4VV20CI047

Under the guidance of:

Latha D U
Assistant Professor
Dept. of
CSE(AI&ML)

CSE (ARTIFICIAL INTELLIGENCE & MACHINE LEARNING)

 VIDYAVARDHAKA COLLEGE OF
ENGINEERING
DEPARTMENT OF

CSE (ARTIFICIAL INTELLIGENCE & MACHINE LEARNING)

CERTIFICATE

Certified that mini project work entitled “Credit Card Fraud Detection”, is a bona fide work
carried out by ARATHI K (4VV20CI009), DEEKSHITA N (4VV20CI014)
NAGAPRAPTHI P S (4VV21CI031), RESHMA R (4VV20CI042) AND SHREYA B
S(4VV20CI047) in partial fulfillment of the requirement for the completion of V semester in
CSE(AI&ML) of Vidyavardhaka College of Engineering during the year 2023-24. It is certified that
all corrections/suggestions indicated for Internal Assessment have been incorporated in the report. The
report has been approved as it satisfies the academic requirements with respect to Mini Project work.

Signature of the Guide Signature of the HOD

Latha D U Dr. Vinutha D C

Assistant Professor Professor & Head

Table of Contents

1. Introduction 4

2. Algorithm 5

3. Code 6

4. Output 7

5. Conclusion 8
 INTRODUCTION

 Pattern recognition is a field within artificial intelligence and machine learning dedicated to the
identification and interpretation of patterns in data. It involves the development of algorithms and
models designed to recognize and categorize patterns, enabling systems to make intelligent
decisions, predictions, or interpretations based on observed data.
 The primary goal of pattern recognition is to create computational models that can autonomously learn
from data and make informed decisions. This process entails recognizing similarities or differences
between instances and deriving generalized insights from the observed patterns. Applications of
pattern recognition are vast, spanning fields like computer vision, speech recognition, natural language
processing, medical diagnosis, and finance.
 The versatility of pattern recognition is evident in its application across various domains. In computer
vision, for example, pattern recognition algorithms excel at identifying objects or faces in images,
while in speech recognition, they enable the understanding and interpretation of spoken language.
Similarly, in medical diagnosis, these algorithms can analyze complex patterns in medical images to aid
in disease detection and classification.
 Patterns come in diverse forms, ranging from spatial patterns in images to temporal patterns in time-
series data. The nature of patterns is highly dependent on the type of data under consideration and
the specific problem being addressed. The ability to recognize and understand these different types
of patterns is a fundamental aspect of pattern recognition.
 The pattern recognition process involves several stages. It starts with the acquisition of data,
followed by the extraction of relevant features that characterize the patterns. Subsequently, these
features are compared with patterns stored in a model or database in the pattern matching stage. The
decision- making phase uses the matched patterns to make informed predictions or decisions. The
process is iterative, involving adaptation and updates to the model based on feedback or new data.
 A variety of techniques are employed in pattern recognition, including statistical methods, machine
learning algorithms (e.g., neural networks, support vector machines, and decision trees), and deep
learning approaches. The selection of a specific technique depends on the complexity of the
problem and the nature of the data being analyzed.
 Pattern recognition faces challenges such as variability in data, noise, and the curse of dimensionality.
Overcoming these challenges necessitates the use of techniques like feature engineering, careful
model selection, and optimization to develop robust and accurate models.
 Pattern recognition has evolved over time, propelled by advancements in computing power, algorithmic
sophistication, and the availability of vast amounts of data. Deep learning, a subset of machine
learning, has particularly demonstrated remarkable success in handling complex patterns, solidifying its
4|Page
 position as a key technology in many pattern recognition applications.

5|Page
 Decision Tree Algorithm
 A decision tree is a powerful and interpretable machine learning algorithm that is widely used for
classification and regression tasks. It is a tree-like model where each internal node represents a decision
based on the value of a specific feature, and each leaf node represents the outcome or prediction.
 Decision trees are popular in various fields due to their simplicity, ease of understanding, and the
ability to handle both categorical and numerical data. This algorithm recursively splits the dataset into
subsets based on the most significant feature at each node, effectively creating a hierarchical structure
that leads to a decision or prediction at the leaves.
 Decision trees are not only valuable for predictive modeling but are also utilized in data exploration,
helping to identify important features and relationships within the data. Their interpretability,
flexibility, and ability to handle complex decision boundaries make decision trees a fundamental
component in the machine learning toolkit.

# Steps involved in KNN Algorithm #

1. Load the data

2. Initialize K to your chosen number of neighbors
3. For each example in the data
4. Calculate the distance between the query example and the current example from the data.
5. Add the distance and the index of the example to an ordered collection
6. Sort the ordered collection of distances and indices from smallest to largest (in ascending order)
by the distances
7. Pick the first K entries from the sorted collection
8. Get the labels of the selected K entries
9. If classification, return the mode of the K labels

6|Page
 CODE SNIPPET

import numpy as np
import pandas as pd
import seaborn as
sns
data=pd.read_csv('card_transdata.csv')
data.head()
data.info()
data.isnull().sum()
sns.heatmap(data.corr(),annot=True)
x=data.iloc[:,:-1]
x.head()
y=data.fraud
y.head()
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
X_train,X_test, y_train, y_test=train_test_split(x,y,test_size=0.20, random_state=10)
classifier = DecisionTreeClassifier()
classifier.fit(X_train, y_train)
from sklearn.metrics import accuracy_score
print("Accuracy is:",accuracy_score(classifier.predict(X_test),y_test))
from sklearn import tree
tree.plot_tree(classifier);

7|Page
 OUTPUT

8|Page
 CONCLUSION

OVERVIEW:
--We did Exploratory data Analysis on the features of this dataset and saw how each feature is
distributed.
--We analysed each variable to check if data is cleaned and normally distributed.
--We cleaned the data and removed NA values.
--We calculated correaltion between independent variables and found that applicant income and loan
approval have significant relation.
--We created dummy variables for constructing the model.
--Finally, we got a model with coapplicant income as independent variable with highest accuracy.
--We tested the data and got the accuracy of 60.9%.
FUTURE SCOPE:
In the near term, the banking software could be more reliable,accurate, and dynamic in nature and can be fit
in with an
automated processing unit.

9|Page