Abstract :With the advent of new Technologies, nowadays electronic gadgets and online shopping’s
more popular. Banking and Online shopping has now become most common activities amongst
the masses. As technology advances so does the risk associate with these transactions. The ease
of use in this online transaction has now become more popular across the world. So it essential
that we need to be very cautious on the increased Fraud activities. Online Fraud is an illegal
activity that can occur when we do electronic transactions. Fraud has increased and created
more risk that has serious financial loss in the financial industry. As a result, these financial
institutions have enforced various techniques to improve their fraud detection methods. Since we
are in the age of Information Technology, Data rules the world. So, Data mining techniques are
widely used to for fraud detection. There are various algorithms such as Anomaly Detection
Algorithm, Decision Tree, Random Forest, K-Nearest Neighbor, K-Means used for fraud
deduction. The type of fraud doesn’t remain the same in each case, so this becomes very crucial
in coming up with the best algorithm for the fraudulent transaction. This paper presents the
survey of those techniques and predicts the best algorithm to detect the fraudulent transaction
based on a given scenario.