K. K. Wagh Institute of Engineering Education and Research, Nashik
Department of Computer Engineering Presentation on Data Mining and Warehousing (DMW) Mini Project
Credit Card Fraud Detection
PID: 26 Guided By: Prof. Dr. S.S. Sane and Prof. D.M. Kanade Group members: DIVISION Roll No. NAME OF STUDENT BE-B 29 Ashutosh Srivastava BE-B 30 Vaibhav Bhavsar BE-B 31 Vaibhav Chalse BE-B 32 Vinay Deokar
Monday, July 05, 2021 Credit Card Fraud Detection 1
PROBLEM STATEMENT
• To recognize/detect fraudulent credit card transactions, so that
customers are not charged for items that they did not purchase
Monday, July 05, 2021 Credit Card Fraud Detection 2
OBJECTIVES
• To recognize/detect fraudulent credit card transactions
• Analyzing the dataset by using data preprocessing techniques and libraries • Analyzing the dataset by plotting various graphs, curves, histograms and visual representations • Applying various machine learning algorithms to improve the efficiency and accuracy score of the machine learning model
Monday, July 05, 2021 Credit Card Fraud Detection 3
OUTCOMES
• A Machine Learning Model to recognize/detect fraudulent credit card
transactions is developed • Implemented 3 ML algorithms for better efficiency and accuracy score • Better accuracy is obtained in Decision Tree Algorithm
Monday, July 05, 2021 Credit Card Fraud Detection 4
DATASET
• The dataset contains transactions made by credit cards in September
2013 by European cardholders (Keggle Dataset) • This dataset presents transactions that occurred in two days, where we have 492 frauds out of 284,807 transactions. The positive class (frauds) account for 0.172% of all transactions
Monday, July 05, 2021 Credit Card Fraud Detection 5
TOOLS & LIBRARIES USED
Tools Libraries
Anaconda Navigator Numpy
Jupyter Notebook Pandas
Language: Python 3.8 Matplotlib
Seaborn
Scikit Learn
Monday, July 05, 2021 Credit Card Fraud Detection 6
ALGORITHMS USED
• KNN (k-nearest neighbors) Algorithm
• Decision Tree • SVM (Support Vector Machine)
Monday, July 05, 2021 Credit Card Fraud Detection 7
BLOCK DIAGRAM
Monday, July 05, 2021 Credit Card Fraud Detection 8
DEMO
Monday, July 05, 2021 Credit Card Fraud Detection 9
ACCURACY COMPARISON
ALGORITHM ACCURACY
KNN (k-nearest neighbors) Algorithm 99.85 %
Decision Tree 99.86 %
Support Vector Machine (SVM) 99.84 %
Decision Tree Algorithm (99.86 %) has slightly better accuracy than KNN (99.85 %) and SVM (99.84 %) Algorithm.
Monday, July 05, 2021 Credit Card Fraud Detection 10
APPLICATIONS • Financial Institutions: Banks/NBFC, AMC, Lending Companies can detect fraudulent credit/debit card transactions • Credit Card Companies: They can detect fraudulent credit/debit card transactions happened using their cards and can immediately block them
Monday, July 05, 2021 Credit Card Fraud Detection 11
CONCLUSION
• A Machine Learning Model to recognize/detect fraudulent credit card
transactions is developed • Successfully implemented 3 ML algorithms for better efficiency • Decision Tree Algorithm (99.86 %) has slightly better accuracy than KNN (99.85 %) and SVM (99.84 %) Algorithm.
Monday, July 05, 2021 Credit Card Fraud Detection 12
THANK YOU!
Monday, July 05, 2021 Credit Card Fraud Detection 13