ATM Fraud Identification Using Machine Learning

DOI

No DOI

Published In

Volume 3, Issue 3 (2022)

Published Online

May-June 2022

Pages

74-77

Article Type

Research Article

Check for updates

No DOI available for cross-reference

Abstract: In today's environment, computer networks are vital for making communication processes more efficient and agile. The increasing requirement for data transmission volume and agility, as well as the ongoing convergence of these operations to the Internet, mandates the employment of larger and more dependable computer networks. As a result, establishing new ways and tools to assist with this management is vital to assuring the quality of services provided. These tools must be efficient and have a low computing cost in order to allow the research of large scale networks. Furthermore, the process of face recognition and identifying problems should be carried out without the involvement of humans, a topic of research known as Autonomic Management. Withdrawal from an ATM (Automated Teller Machine) system using LINK technology and a card reading device two step of authentications for security face recognitions and otp checking. Withdrawal from an ATM (Automated Teller Machine) system using LINK technology and a card reading device. An attacker can simply conduct fraudulent withdrawals if he obtains an ATM card and pin number. As a result, we recommend a system that incorporates an ATM card scanning system as well as a LINK system. By scanning his card, this individual can gain access to the system. However, once the user has finished the authentication process, he can view the data, but if he chooses the money withdrawal option, he will be requested to enter the LINK machine learning for linear regressions algorithm. This strategy decreases analysis complexity while enhancing network stability and accessibility, allowing administrators to swiftly spot defects, threats, and system failures. As a result, this study employs a variety of traffic characterization models, including the Digital Signature of Network Segment utilizing Flow Analysis (DSNSF), the (ARIMA) approach, and Principal Component Analysis (PCA). Proposed Map Reduce Based Holt Winters Method and Principal Component Analysis (PCA). The DSNSFs derived by the proposed models are compared to real-world traffic of bits and packets in a real network environment, and then submitted to particular tests to determine their accuracy. The suggested approach produces a workable solution that is a significant advance over current systems.