Data-Driven QoE Model for Addis Ababa LTE Video Streaming Using Fuzzy Logic Inference System

Nowadays, the video streaming services become the most dominant service as people are more interesting watching online television programs and Video on Demand (VoD). This requires a high speed and high capacity network infrastructure. The Long Term Evolution (LTE) network infrastructure of Addis Ababa has faced this network capacity and data rate demand in order to have a good Quality of Experience (QoE). To takeover this challenge, Ethio Telecom should have appropriate assessment methodology for QoE. This thesis outlines the means of QoE modelling issue to measure QoE objectively. It proposes a QoE model which is used to measure the QoE from Quality of Service (QoS) parameters using Fuzzy Logic Inference System (FIS). This has been done by collecting end-to-end QoS data from network management system which is used as input for the model and the proposed model has been validated using a dataset collected from customer survey. The Model developed is essential for replacing the subjective measurement techniques which are costly and inefficient being influential to user context, terminal characteristics, application software, user capability, etc. In addition, the model developed is helpful for business decision making, network planning, optimization and operational support activities in manual systems or system based for self-organizing networks according to the infrastructures implemented. The result of correlation, regression, and Four-way ANOVA show that the Stall Frequency (SF) and the Start Delay (SD) plays the major impact on the LTE video streaming QoE by 33% and 25% respectively. The validation result shows the proposed model is an accurate, consistent and linear compared to the existing models.