2. INTRODUCTION
• Now a day the usage of credit cards has dramatically
increased.
• As credit card becomes the most popular mode of
payment for both online as well as regular purchase,
cases of fraud associated with it are also rising.
• Major Risk – credit card detail is known to other
2
3. Machine Learning
• Machine learning is a method of data analysis that automates
analytical model building.
• The concept behind using machine learning for fraud
detection is that fraudulent transactions have specific
features that legitimate transactions do not.
• Based on this assumption, machine learning algorithms
detect patterns in financial operations and decide whether a
given transaction is legitimate
3
4. ABSTRACT
• In this project, machine learning algorithms are used to
detect credit card fraud.
• Standard models are firstly used.
• Now used hybrid model used.
• Our model is BiLSTM- MaxPooling-BiGRUMaxPooling
which based on bidirectional Long short-term memory
(BiLSTM) and bidirectional Gated recurrent unit (BiGRU).
4
5. • We also applied six machine learning classifiers which are:
Naïve base, Voting, Ada boosting, Random Forest, Decision
Tree, and Logistic Regression.
• To evaluate the model efficacy, a publicly available credit
card data set is used.
• Then, a real-world credit card data set from a financial
institution is analyzed.
• The experimental results positively indicate that the
majority voting method achieves good accuracy rates in
detecting fraud cases in credit cards.
5
6. Problem statement
• Credit card fraud is a serious problem in financial
services.
• Billions of dollars are lost due to credit card fraud
every year.
• Online payment does not require physical card
• Anyone who know the details of card can make fraud
transactions
• Currently, card holder comes to know only after the
fraud transaction is carried out.
• No mechanism to track the fraud transaction
6
7. Disadvantages
• The high amount of losses due to fraud.
• Testing credit card FDSs using real data set is a
difficult task.
• The fraud has to be deducted in real time and the
number of false alert.
7
8. Advantages
• The system is very fast due to AdaBoost
Technique.
• Effective Majority Voting techniques.
• Good accuracy rates in detecting fraud.
.
8
9. Algorithm
• Machine Learning Algorithm
– They are used in conjunction with the AdaBoost
and majority voting methods.
– The Decision Tree (DT) is a collection of nodes
that creates decision on features connected to
certain classes.
– TheRandom Tree (RT) operates as a DT operator,
New nodes are established until the stopping
criterion is met.
9
10. Forming Hybrid Models
• Adaptive Boosting or AdaBoost is used in conjunction with different
types of algorithms to improve their performance.
• AdaBoost weak learners in favor of misclassified data samples.
• Majority voting is frequently used in data classification, which
involves a combined model with at least two algorithms.
• Each algorithm makes its own prediction for every test sample.
• The final output is for the one that receives the majority of the votes.
10
13. Conclusion
• A study on credit card fraud detection using machine learning algorithms has been presented
in this project.
• we have performed several machine and deep learning models to detect whether an online
transaction is legitimate or fraud on the IEEE-CIS Fraud Detection dataset as well built our
model which is BiLSTM-MaxPooling-BiGRUFigure MaxPooling that based on bidirectional
LSTM and GRU.
• The results from machine learning classifiers show that the best AUC(Area Under the
Receiver Operating Characteristic Curve)was 80% and 81% that achieved by hard voting
with undersampling and oversampling technique.
• The results from machine learning classifiers were not promising compared with our model
that achieved 91.37% AUC.
13