Supervised Machine Learning based Classification of Video Traffic Types

Abstract

The main topic of the article is accuracy evaluation of supervised machine learning algorithms performance applied to real network traffic data. The main task to be solved by supervised learning is classification of video traffic type - streaming (real-time) video or on-demand video (a record). The experiment has been performed for the same video fragment with data filtering and without it. The results have been summarized in form of tables with accuracy assessment for multiple commonly used supervised machine learning algorithms.