Enhancing IoT Wi-Fi Networks Through QoS-QoE Correlation: A Random Forest Approach for Predictive Mean Opinion Score Modeling

The Internet of Things (IoT) domain has garnered considerable attention from the scholarly community, with Quality of Service (QoS) emerging as a pivotal concern within IoT frameworks. Furthermore, the perceptions and satisfactions of end-users within the IoT milieu hold substantial weight, underscoring the necessity for rigorous Quality of Experience (QoE) assessments. Therefore, this work suggests a QoE scheme for IoT-enabled devices that should adopt the QoS technique, which would position the Wi-Fi end-use case more efficiently. In this scheme, the proposed scheme identifies key data elements related to Wi-Fi devices, including the environment, which would then be fed into classical classification algorithms (such as Linear regression, Decision tree and Random Forest) to provide us with the Mean Opinion Score (MOS is user experience ranging from 1 through 5 where 5 is the best and 1 is the worst experience). MOS will then be utilised to tune Wi-Fi characteristics to improve the overall user experience. This scheme will involve studying various types of parameters related to QoE (such as Signal strength, Noise, Data rate, Packet loss, Latency and Jitter) followed by an experiment. This will eventually be adapted to the existing Wi-Fi solution to improve the throughput, delay, jitter, and other network parameters to provide a better user experience.