Detect Legitimate and Illicit Transactions Using Machine Learning

The advancement of technology and E-Commerce has correspondingly increased the use of online transactions for almost every activity. However, this rapid growth in payment modes also increased fraud associated with the online transactions. Many particulars have lost their amount due to such frauds and this experience resist them to use online method of payment for future operations. Money laundering is one example of financial fraud, which is recognized as a significant criminal process that involves using cash gained via illegal means for terrorist or other illegal activities. Because these illicit activities entail intricate networks of commerce and financial transactions, it is challenging to identify the fraud firms and identify their characteristics. Algorithms of machine learning have become important tools in the battle against online transaction fraud. These algorithms examine huge amounts of transactional data and spot patterns suspicious of fraudulent activity by utilizing the modern technique and computational approaches. Therefore, it is important to detect those frauds. There are many Machine Learning algorithms which are evolved in detection of various fraudulent transactions. In this paper we apply XGB Classifier which is the boosted version of decision tree and compare performance, accuracy and precision with various supervised machine learning algorithms.

Перевод пока недоступен