Loan Fraud Detection using Machine Learning Algorithm

Abstract

The widespread usage of the Web has had a massive effect on the growth of online card transactions, especially at the beginning of the previous few years. Because of the increase in internet transactions, the global banking system has been forced to deal with or confront an unexpected number of fraudulent operations. As a reason, rule-based algorithms were created to detect high-risk transactions and allow specialists to authenticate whether or not they were fraudulent. The present intruders used the static nature of rule-based systems as a defensive measure to avoid detection. As a result, researchers set out to develop adaptive fraud detection systems based mostly on machine learning techniques, including deep learning, which is a relatively new application. The widespread use of the Internet, notably at the start of the previous decade, had a considerable impact on the rise in online card transactions. The rise of internet transactions has forced the worldwide banking system to deal with or confront an unexpected amount of fraudulent operations. As a result, rule-based algorithms were developed to identify high-risk transactions and allow specialists to confirm whether they were genuine or not. To avoid detection, the current attackers used the static nature of rule-based systems as a countermeasure. As a result, researchers set out to develop adaptive fraud detection systems based mostly on machine learning techniques, including deep learning, which is a relatively new application.