Abstract

Credit Card Fraud Detection Using Machine Learning Algorithm is a study of customer behaviour in bank transaction. In this we are taking an original bank transaction dataset, and dividing it to three different levels of fraud rate. We apply various machine learning algorithms to these three versions and identify the fraud rate and record the same. It reports the accuracy of each algorithm and how it varies as the fraud rate increases. It also helps us to know which attributes must be considered and necessary for identifying the fraud in transactions. This project gives the insight about three machine learning algorithms that is, Support vector machines, Local Outliers and Isolation forest for credit card fraud detection. These three algorithms are applied to three different datasets with increasing fraud rates and their accuracies have been recorded via graphs and charts.