Detecting Fraud Transaction using Ripper Algorithm Combines with Ensemble Learning Model

In the context of 4.0 technology revolution, which develops and applies strongly in many fields, banking sector is considered to be leading one, application algorithms detect fraud extremely important. necessary. recent years, credit card transactions including physical payments online have become increasingly popular countries around world. This convenient payment method attracts more criminals, especially fraud. As a result, banks world developed detection prevention systems for each transaction. Data mining one techniques applied these systems. study uses Ripper algorithm fraudulent on large data sets, results obtained with accuracy, recall, F1 measure than 97%. research then used combined Ensemble Learning models transactions, are 99% reliable. Specifically, this model using Gradient Boosting has improved predictive ability very reliable results. The use machine learning expected new topic will widely banks’ or organizations’ activities related e-commerce.