IoT devices analysis using neural networks ensemble trained on unbalanced sample

An approach to identifying anomalous situations in network segments of the Internet of Things based on an ensemble of classifiers is considered. Classifying algorithms are tuned for different types of events and anomalies using training samples of different composition. The use of an ensemble of algorithms makes it possible to increase the accuracy of the results due to collective voting. The experiment performed using three neural networks identical in architecture is described. The results of the assessment were obtained both for each classifier separately and with the use of an ensemble