Convolutional neural network for detecting anomalies in the control system of a machine-building enterprise

Abstract

The article is devoted to the study of the possibility of using machine learning methods and the apparatus of neural networks to ensure a given performance and security of the enterprise infrastructure. Continuous monitoring of computer networks and enterprise resource planning systems determines the stability of the enterprise. The infrastructure of any enterprise is gradually expanding and increasing, not only due to the growth of the geography of networks, but also due to the use of equipment from various manufacturers. Therefore, it is important to create intrusion detection and prevention systems with automated monitoring of equipment, processes and user activity. The article describes the use of convolutional neural networks for predicting indicators and technical characteristics of equipment and communication channels. An anomaly detection module structure has been developed for monitoring the operation of the enterprise resource planning system. An example of training data for a neural network is given. The experimental results confirmed the effectiveness of the proposed approach.