E-Commerce Fraud Detection Using Support Vector Machine and Naïve Bayes Algorithm

Abstract: The prevalence of online fraud cases, including e-commerce fraud, is rising as a result of technological advancements and the speed with which cybercriminals can change their methods of operation. Scams are nothing new, but as the frequency of transactions without currency rises, so does the trend of online fraud. People are purchasing more goods online as a result of the COVID-19 quarantine because they want to be safe or because the items, they require are hard to get in the shuttered local stores. The best course of action in this circumstance is to implement a fraud prevention service that automatically identifies fraudulent behaviour patterns, associated with the time, place, and device name associated with the login or transaction. This will prevent fraudsters from using the data they stole. You can halt fraudsters before they start a transaction by spotting suspicious activity on an account. Through relevant historical data from databases and machine learning techniques, this study aims to identify fraud patterns in e-commerce transactions. Based on email, payment methods, payment method providers, and transaction volume, this research will train a computer or system that can predict fraud patterns. Machine learning must be used to improve fraud protection in e-commerce since it allows machines to be analysed using learning algorithms. Support vector machine and naive Bayes will be the algorithms employ.

Keywords: Machine Learning, Support Vector Machine, Naïve Bayes, Classification, Fraud E-Commerce.