Qoe Based Service Quality Evaluation In Mobile Networks

Today’s telecommunication market has stepped into an era in which new services are rapidly developed and deployed. Facing serious homogenization of services, Communication Service Providers (CSPs) are eagerly to research and implement Customer Experience Management (CEM) programs in order to differentiate themselves, maintain their advantages over their counterparts and gain profits. And as a critical part of CEM program, the quality of delivered service still represents the core competence of a CSP. CSPs need to be customer-experience-focused rather than network-focused, and to enhance their abilities to monitor and evaluate the delivered quality of digital services in the perspective of the customers. Hence, how to efficiently monitor and evaluate the customer experience of the services is an urgent task for the CSPs.The thesis first studies in detail the theory and technology of CEM of telecommunication industry. By analyzing the customer experience concept and metric model, specific signaling and business processes, the thesis establishes a hierarchy of perception index of WCDMA voice service. Then the mapping algorithm of the indices is researched and their weight relationships within the hierarchy are calculated.And by the research and analysis of four widely used multi-criteria evaluation approaches, the thesis recommends that neural network technique would be more appropriate for telecom customer satisfaction evaluation. Thus, a QoE perception index system based customer satisfaction evaluation approach which applies the BP neural network technique is proposed. Taking the data of the index system as input, and utilizing the neural network’s ability of learning and generalization, this method can dynamically evaluate the whole customer satisfaction. And by experiment, the evaluation proves to be rather accurate, which would be helpful and applicable for the CSPs.