A Comprehensive Review of QoE Prediction Models in Dynamic Edge Networks: Challenges and Hybrid Approaches

The rapid expansion of dynamic edge networks has revolutionized time-sensitive applications, including precision agriculture, UAV-based emergency systems, and ۶G edge intelligence. However, accurately predicting user Quality of Experience (QoE) in these environments remains a complex challenge due to fluctuating topologies and temporal variations. Traditional models often emphasize isolated aspects, neglecting integrated structural and temporal interactions. This review systematically examines QoE prediction models, categorizing them into deep learning-based, reinforcement learning-driven, and feature-oriented approaches. It provides a detailed analysis of their strengths, limitations, and performance metrics such as accuracy and latency. The study highlights the potential of hybrid frameworks to address gaps in scalability and real-time adaptability, offering guidelines for advancing QoE optimization in intelligent edge ecosystems.

کلیدواژه ها:

Quality of Experience (QoE) Prediction ، Dynamic Edge Networks ، Hybrid Approaches ، Graph-Temporal Analysis ، QoE Datasets