A new three-layer QoE modeling method for HTTP video streaming over wireless networks

Abstract

Quality of experience (QoE) indicates users' perception of an Internet service or business in objective conditions, which is significantly meaningful to service providers or network operators to improve their quality of services and to utilize valuable network resources efficiently. Online video service has become the most fast-growing Internet service in recent years. Building models for QoE evaluation of video services has become a hot research topic. In this paper, we propose three metrics to evaluate buffering situations of HTTP video streaming, which have been verified to be the main factors that influence the QoE of network videos, and together with three fundamental quality of service (QoS) parameters, to build a three-layer QoE model for HTTP video streaming. Then, we use linear regression and BPNN (back propagation neural network) to reflect the mapping relations of the QoS parameters, buffering-related metrics and QoE, and demonstrate that BPNN performs better than linear regression in this situation.

Perception of quality in cloud computing based services

“…The scenario considers that the learner is using an online learning algorithm that builds a prediction model from this data, while prediction model is used to predicting the QoE from QoS available data. The linear regression and BPNN (back propagation neural network) have been used to reflect the mapping relations of the QoS parameters and QoE, in order to build a three-layer QoE model for HTTP video streaming [11]. Another acceptable QoE model is based on the results of comprehensive user studies on subjective quality acceptance assessments for mobile video services [12].…”

Online QoE assessment model based on incremental stacked multiclass classifier

“…For example, in [10] an application of various ML techniques for QoE estimation was debated. Other research works have employed a neural network model (NN) to assess perceptual video quality [11]. Moreover, the Decision Tree (DT) model was used in [12] to build a No reference objective model to estimate Enduser QoE of video stream services.…”