Abstract
Learning more about the customers' perceived quality is of significant importance for service providers. This information can be used to better manage the system resources and provide more value to the users. However, it is often not easy to develop autonomic mechanisms that can accurately measure the delivered quality of multimedia services due to the subjectiveness of the perceived quality. While the most accurate methods involve subjective testing, which is costly and impractical, the existing objective ones lack of accuracy. In this work we present a methodology that uses limited initial subjective tests to develop prediction models for Quality of Experience (QoE) as perceived by the viewers. These prediction models are further used as part of a QoE estimation platform that assesses the value of perceived quality based on the underlying conditions of the system. © 2010 IEEE.