Paper Title
Comparative Study on Credit Card Fraud Detection
Abstract
Credit card fraud has become one of the growing problems. The Credit Card Fraud stands as major problem for
worldwide financial institutions.It's crucial that credit card companies can classify fraudulent activities so that consumers
don't have to pay for items they didn't agree to. With the making number of clients, credit card fakes are in addition reaching
out at a comparative rate. The charge card data of a specific individual can be theft unlawfully and can be utilized
unfairly.The credit card information of a particular individual can be robbery unlawfully and can be used wrongfully.. In this
paper we have discussed different extortion location models which check if another transaction is fraudulent.
Machine learning is widely regarded as one of the most effective methods for detecting fraud. Developing effective fraud
prevention methodologies is critical that’s why more and more advanced machine learning techniques are used to reduce
these disadvantages and assist fraud investigators. However there are number of challenges appear to tackle this problem,
such as lack of publicly available data due to confidentiality reason,imbalanced class sizes,variant fraudulent behavior. To
reduce imbalanced class sizes, we use smote sampling technique.In this work a comparison of five supervised machine
learning algorithms :logistic regression, naive bayes, KNN ,Random Forest, Decision Tree to separate among certified and
fraudulent transactions execution of the algorithms is assessed dependent on ordinarily acknowledged metric for example
accuracy, precision and sensitivity.
Keywords - Credit Card, Credit Card Fraud, Credit card, Machine Learning, Supervised, SMOTE