Abstract
Constant improvement of information technology leads to expansion of sizes of local networks of enterprises and issues connected to its organization and monitoring.Considering that every minute a huge amount of heterogeneous data is generated within the network, operating a computer network LAN becomes a nontrivial task.Since a computer network consists of a wide variety of devices (switches, servers, routers, computers etc.), and each of these devices utilizes a high number of application software, which is a source of immense amount of information, the forecast based on processing and analysis of this data would be best made by an algorithm of machine learning.The presence of a sufficiently accurate prediction of malfunction will be able to ensure high reliability, safety and economic efficiency of operating the equipment of local area networks segments.This article proposes a model for predicting equipment failures in a particular case based on the Random Forest machine learning algorithm.The key stages of building and tuning the model are considered.The model includes several sub-models predicting equipment failure using actual and predicted readings from sensors.