Client-based QoS data selection and modeling using generalized extreme value theorem and linear opinion pool

Assuring Quality-of-Service (QoS) guarantees to mobile clients is still a long-standing problem in today´s wireless mobile networks. In this paper, we propose new algorithms that validate the accuracy of service measurements collected from the mobile clients to construct a precise QoS model. The proposed algorithms utilize the Linear Opinion Pool (LOP) approach in conjunction with Generalize Extreme Value theorem (GEV) to converge to a precise QoS model. The main objective of this work is to construct a QoS model that excludes the out-of-profile data that is collected from the Mobile Clients (MCs). Therefore, any MC with unreliable data is considered as un-trusted. The proposed approach is effective in providing service providers with a better assessment tool to evaluate and improve their services. The results show the usefulness of our algorithms and their ability to recognize and exclude the data collected from un-trusted MCs; thus, creating a precise QoS model.