An Empirical Technique to Establish the Ideal Parameters for an Adaptive Neuro-Fuzzy-Based Qoe Optimized LAN Model

The expansion of real-time applications like video streaming, conferencing, and interactive services has made Quality of Experience (QoE) optimization in communication networks more crucial. An adaptive neuro-fuzzy-based QoE optimization model for local area networks (LANs) was developed and evaluated in this research. Multiple predictor combinations were assessed using experimentally generated throughput, delay, jitter, packet loss, and bandwidth datasets. The study finds that throughput, delay, jitter, and packet loss are the best predictors of network QoE using regression analysis and evaluation metrics including the Akaike Information Criterion (AIC), Adjusted R2, and Root Mean Squared Error (RMSE). Superior prediction accuracy was demonstrated by the chosen four variable model, which maintained strong explanatory power while achieving the lowest AIC of 12062.53 and RMSE of 100.12.This work incorporates a Neuro-Fuzzy framework in its experiment and analysis in contrast to traditional methods of regression exclusively. These results show how useful Neuro-Fuzzy modeling is for identifying nonlinear relationships between end user experience and network performance, providing a scalable framework for real-time QoE control in LAN settings. In order to establish a fusion of objective–subjective framework, future work will expand the framework to diverse networks, explore rule-reduction schemes to advance scalability, and incorporate more subjective QoE measurements. A strong and understandable AI-driven method for QoE optimization in network performance management is provided by this study.