1. Intelligent Transaction System for Fraud Detection using Deep Learning Networks
- Author
-
N M Sanjeev Kumaar, K Sakthi Eswaran, J Fenila Naomi, and R Roshan Jeniel
- Subjects
History ,business.industry ,Computer science ,media_common.quotation_subject ,Deep learning ,Payment ,Computer security ,computer.software_genre ,Computer Science Applications ,Education ,Consistency (database systems) ,Credit card ,Issuer ,Order (business) ,Cash ,Artificial intelligence ,business ,Database transaction ,computer ,media_common - Abstract
Detecting online transaction fraud is a basic study of the new era of electronic transactions. Because the payment patterns of customers and the fraud behaviour of offenders are continually changing, improving the consistency of the fraud detection model and ensuring its stability is exceedingly challenging. In this report, we will look at We concentrate on acquiring deep feature representations of legal and fraud transactions from the perspective of a deep neural network’s loss function in this report. Our aim is to increase the separability and discrimination of features in order to boost the efficiency and stability of our fraud detection platform, with the rapid evolution of the technology, the world is turning to use online transaction instead of cash in their daily life, which opens the door to many new ways for fraudsters to use these cards in a nefarious manner. Global losses are projected to reach $35 billion by 2020, according to the Nilson report. To guarantee that users of these credit cards are secure, the credit card issuer should provide a program that protects them from any threats they can experience. As a result, we illustrate our framework for predicting whether transactions are genuine or illegitimate using Kaggel’s IEEE-CIS Fraud Detection dataset. BiLSTM-MaxPooling-BiGRUM is the name of our model. Long bi-directional gated repeated unit and long bi-directional memory term (BiLSTM) are used in axPooling (BiGRU).
- Published
- 2021