Algorithms for proactive security of industrial systems based on machine learning technologies

Abstract

Approaches to improving the efficiency of network attack detection algorithms in heterogeneous industrial networks based on machine learning technologies are considered. An algorithm for analyzing and processing network traffic has been developed in the task of detecting malicious network activity. The Electra dataset is used to train the proposed machine learning models and heterogeneous neural network models.

Алгоритм анализа и обработки сетевого трафика был разработан для решения задачи обнаружения вредоноснойсетевой активности. Набор данных Electra используется для обучения предложенных моделей машинного обучения и гетерогенных нейронных сетевых моделей.