IPTV QoE evaluation method based on data-mining

As it is difficult for service providers to evaluate the QoE of IPTV customers, and massive network operation data could not be used to solve QoE problems, through the analysis of IPTV network operation data, metrics, current IPTV QoE evaluation methods and relevant data mining algorithms, a data-mining based IPTV QoE evaluation method was proposed to build a QoE model from operation data. The main steps of the method included metrics correlation analysis, feature seletion, dimension reduction, regression analysis, and QoE alert. The results of experiments on real data sets show that the method can reach alert accuracy of 66.35%, when QoE score is set at 80 as a threshold. © 2016, Editorial Board of Journal of Huazhong University of Science and Technology. All right reserved.