eva‎luating quality of experience for social media applications in cellular networks using a machine learning approach

2/11/2026 12:00:00 AM

This research presents a machine learning-based approach capable of calculating the Quality of Experience (QoE) of social media under various cellular network conditions. Since QoE is a user-centered metric, its calculation is a challenging and time-consuming task. The dependency of QoE on network parameters such as Quality of Service (QoS) can help simplify its calculation. In this thesis, we first review the general concepts and previous work done in the field of QoE eva‎luation. Then, we present an innovative approach that can map cellular network parameters to QoE with appropriate accuracy. The proposed approach uses four cellular network parameters: RSSI, SINR, RSRP and RSRQ along with the P.1203 standard to estimate QoE. The innovation of this research is that this approach can calculate QoE without the need for surveys or the presence of real users. In fact, we simulate a conversation between two users on social media by implementing two different software applications. In this approach, we will map the mentioned parameters to QoE using machine learning algorithms. Finally, by analyzing the resulting machine learning models, the importance and impact of the mentioned parameters on QoE are examined. Our proposed model was able to calculate QoE in different network conditions with acceptable accuracy. According to the results from the analysis of the machine learning models’ outputs, the SINR parameter has a significant effect in estimating the Quality of Experience.

كيفيت تجربه , كيفيت خدمت , رسانه‌هاي اجتماعي , شبكه‌هاي تلفن همراه

QoE , QoS , Social networks , Mobile networks

Mohammad Hosseinali

Abolfazl Diyanat