Quality of Experience (QoE) and Network Performance Modelling for Multimedia Traffic
Table 5.
QoEestimation model
|
| Model |
RMSE |
R² Score |
Computation Complexity |
Observations |
|
|
| Exponential Mapping |
0.48 |
0.81 |
Low |
Captures rapid QoE decline at early QoS degradation but underestimates recovery at low loss. |
|
| Logistic Model |
0.42 |
0.84 |
Low |
Models saturation behavior accurately but less adaptable across scenarios. |
|
| Random Forest Regression |
0.25 |
0.92 |
Medium |
Provides robust prediction but needs large training data. |
|
| Neural Network Model |
0.22 |
0.95 |
High |
Best prediction accuracy; effectively models nonlinearities. |
|
| Proposed Hybrid Model |
0.19 |
0.97 |
Moderate |
Achieves optimal trade-off between accuracy and complexity. |
|
|
|
