Quality Prediction of Mobile Video Service Based on Radio Access Network Log Data

Quality of Experience (QoE) of mobile video is being raised as the most important parameter for network operators due to the rapid increase in mobile video traffic. But in the mobile network, the network-level Quality of Service (QoS) is provided and this network-level QoS does not exactly represent the user-level QoE. In this paper, we propose a system with two modules for forecasting mobile video quality based on network Radio Frequency (RF) metrics that are related to QoS. We develop a Network Quality Prediction (NQP) module which can analyze patterns of RF metrics and predict the next RF metrics patterns. In addition, we design a Video Quality Scoring (VQS) module which can predict Video Mean Opinion Score (VMOS) by mapping predicted RF metrics to VMOS. The proposed system, NQP and VQS, shows a meaningful prediction performance of VMOS.