Prediction of States of Information and Communication Systems using Machine Learning

Today, new realities dictate entirely different working conditions for telecommunications systems. Telecommunications systems now generate large volumes of heterogeneous traffic, which is why classical methods of information management and processing face certain challenges when operating efficiently in real-time mode. For example, modern mobile networks exhibit complex behavior due to network heterogeneity and a large number of diverse mobile devices or systems operating on mobile networks. Proactive monitoring, based on machine learning, aims to address this issue through the analysis of big data and the discovery of new behavior patterns. This article describes the existing problem in monitoring systems and the interim results of using machine learning as an additional tool for predicting the behavior of telecommunications networks. Initially, we will discuss the key concepts related to machine learning. Then, we will present a conceptual framework for a machine learning-based prediction system. We will review interim results on test data and examine the challenges of implementing such systems in practice.