Quality of Experience (QoE) and Network Performance Modelling for Multimedia Traffic
Table 2.
Comparison of QoE Assessment Methods
|
| Method Type |
Technique / Example |
Measurement Basis |
Advantages |
Limitations |
Typical Use Case |
|
|
| Subjective |
Mean Opinion Score (MOS), ITU-T P.800, DSCQS |
Human perception and user ratings |
High accuracy, directly reflects user perception |
Costly, time-consuming, not scalable |
Laboratory testing, service validation |
|
| Objective (Signal-Based) |
PSNR, SSIM, VMAF |
Comparison of original and transmitted signals |
Automated, reproducible, quick analysis |
Ignore human perception nuances |
Video streaming quality benchmarking |
|
| Objective (Parametric / Hybrid) |
ITU-T P.1203, E-model |
Uses network and codec parameters to infer QoE |
Real-time estimation, scalable |
Requires calibration, may vary by scenario |
Network performance monitoring, adaptive streaming |
|
| Data-Driven (Machine Learning) |
Random Forest, SVM, Deep Neural Networks |
Predicts QoE from QoS datasets |
Adaptive, captures nonlinear relations |
Needs large training data, model interpretability issues |
Intelligent QoE prediction, self-optimizing networks |
|
|
|
