Application of a Mixed Effects Model in Predicting Quality of Experience in World Wide Web Services

This paper investigates the applicability of a mixed effects model in predicting Quality of Experience (QoE) in World Wide Web based multi-media services. An analysis is presented on objective factors and a human factor that may impact on outcomes of observations as fixed effects, population mean, and random effects, group specific, and account for a correlation structure of variations amongst users. The objective metric for QoE evaluation has been based on networking perspectives from the World Wide Web and metrics based on the application layer and human perception. The subjective metric is related to a state of mind factor and the variance of each individual amongst group. The model is developed to account these factors and other potential covariates to QoE assessment during the course of the experiment. The model demonstrates an acceptable correlation between values of the fitted analytical model and observed user scores.