A Model For Mapping QoS To QoE In Mobile Network

Quality of Experience (QoE) is a new metric in wireless network communication, depending on the underlying QoS along with a person’s preferences and the environment surrounding that person, which defines the overall subjective feeling of a user during the interaction with a particular service or application under a certain context. QoE directly reflects the acceptance of a service by combining all the factors from service, network and user. QoE measurement and prediction are complex tasks as they involves a large parameter space such as delay, jitter, packet loss and attributes related with the context and user. These tasks necessitate the development of practical context-aware QoE models that efficiently determine relationships between context and QoE states.In this paper, we propose and validate a novel decision-utility approach for QoE modeling, measurement and prediction. We solve the challenge of QoE measurement and prediction where each QoE states can be measured on a different scale and different units of measurement using a bipolar scale; integrate utility theory and Bayesian networks to realize context-aware QoE measurement and prediction.We validate our models using experimentation and user studies. We prove the correctness of QoE measurement modeling by comparing the outcomes from model calculation and user studies under different user context. For QoE prediction, an overall accuracy of above98%was achieved using10-fold cross validation using three different kinds of Bayesian networks.