Machine learning based QoE prediction in SDN networks

Our work in this paper presents a prediction of quality of experience based on full reference parametric (SSIM, VQM) and application metrics (resolution, bit rate, frame rate) in SDN networks. First, we used DCR (Degradation Category Rating) as subjective method to build the training model and validation, this method is based on not only the quality of received video but also the original video but all subjective methods are too expensive, don't take place in real time and takes much time for example our method takes three hours to determine the average MOS (Mean Opinion Score). That's why we proposed novel method based on machine learning algorithms to obtain the quality of experience in an objective manner. Previous researches in this field help us to use four algorithms: Decision Tree (DT), Neural Network, K nearest neighbors KNN and Random Forest RF thanks to their efficiency. We have used two metrics recommended by VQEG group to assess the best algorithm: Pearson correlation coefficient r and Root-Mean-Square-Error RMSE. The last part of the paper describes environment based on: Weka to analyze ML algorithms, MSU tool to calculate SSIM and VQM and Mininet for the SDN simulation.