Machine Learning for QoE and QoS Control of Slices in a Wide Area Network Test Bed

Abstract

In this work an experimental investigation is reported about the use of machine learning, based both on a regressive approach and on artificial neural network, to evaluate the quality of experience from quality of service and other network measurements as packet losses, delays and traffic congestions, to control the performance of slices defined inside a wide area network test bed. Such a method allows the network to recover the best performance according to a knowledge defined network approach.