Neural Network-Based Risk Assessment for Cybersecurity in Big Data-Oriented ERP Infrastructures

Abstract

Neural networks have significantly evolved in many fields in the last two decades. In particular, neural networks have had a growing impact on risk assessment in cybersecurity systems. Thus, this paper purports to extend human knowledge and intuition by presenting a neural network-based approach for assessing cybersecurity risks. The area of interest in the current research is not related to any system featuring an enterprise resource planning (ERP) in and of itself; it is rather directed to the creation of a perspective featuring the peculiarities and challenges of Big Data-oriented ERP, which is currently being hyped. Security considerations for BDOG thus have to consider the ERP data structure, possibilities, and inconveniences; this shows that BDOG infrastructure, as per current knowledge, has arisen now to be a contemporary factor.

This research aims to create a cybersecurity risk assessment framework invented after predicting and managing challenges in tackling cybersecurity for the BDOG-ERP infrastructure by several of the newest AI advances for comprehensive Big Data access. In particular, in this paper, a new program framework to show the decision-maker the robustness of the cybersecurity risk profile assessment. The conceptual combination of a multilayer perceptron neural network and a network of neural networks that is multilayer will be considered responsible for this model. It is shown that these are the enabling system tools to offer distributed computing and the capability to manage and manipulate significant, interdisciplinary ambiguity and stochastic parameters, factors, and vectors (both systematically and randomly organized). The research illustrates that no risk component is evaluated individually, but is derived from various aspects representing specific components of risk; it is also tested by 60 different scenarios.