A Novel Approach for Credit Card Fraud Detection Using Machine Learning Paradigm

Abstract

The brisk help in online based transactional activities raises the phony cases wherever all through
the world and makes tremendous disasters the individuals and cash related industry. Credit card fraud is
additionally developing alongside the improvement in innovation. It can likewise be said that financial fraud is
radically expanding in the worldwide correspondence improvement. Present day credit card deception is a
significant stress for most by far of the cash related organizations. Online trades become essential, basic and
profitable as a consequence of the formation of credit card. But it makes odds of making transactions fraud by
the fraudster. In order to keep up a vital separation from more fraud, there may be two concept used. One is to
find the fraud, also called fraud identification and other is fraud prevention [1]. In classification of fraudulent
and genuine transaction made by credit card holder perhaps one of the best test beds for the machine learning
and deep learning algorithm. In fact this task itself consists of challenges like data imbalance, concept drift, and
small disjuncts etc. In order to tackle the scenario an ensemble based machine learning model, Logistic
Regression and support vector machine are designed and identify the fraudulent transaction once transaction is
being made. The proposed model outperforms from the literature in terms of measure of MCC is 0.894.