User Satisfaction Prediction Framework for Personalized Wireless Networks: A DNN Approach

Abstract

New data applications, smart devices, and technologies are emerging every year. Future wireless networks will be expected to maintain user expectations in spite of the difficulty of managing explosive increases in network data traffic. Yet user experience is highly subjective and depends on the highly dynamic user satisfaction behavior in the network. For this reason, real-time data-driven user experience modeling and prediction are more relevant than mathematical modeling. Modeling and predicting user satisfaction in real-time will enable wireless networks to make more personalized decisions, which can increase efficiency and user satisfaction. In this paper, we propose a framework for implementing a real-time big data-driven satisfaction monitoring and prediction system for personalized wireless networks based on a dynamic user satisfaction model. Then, inspired by the success of neural networks and deep learning techniques, we implement the proposed framework using a tuned DNN model. Finally, the results of our experiment show the feasibility and preeminence of the proposed framework.