Advances, Systems and Applications
Table 6 Comparative analysis of RDAER vs. existing works
From: An integrated SDN framework for early detection of DDoS attacks in cloud computing
Model | Acc | Pr | Re | F1 score | AUC | Limitations |
|---|---|---|---|---|---|---|
ARIMA [20] | 98.82 | 98.46 | 0.8 | 0.98 | 0.99 | It was less accurate. It only employs one hidden layer |
V-ELM [49] | 99.18 | NA | 99.5 | NA | NA | There is potential to improve accuracy |
LKM-OFS [50] | 96.54 | 84.89 | 89.92 | 85.66 | NA | The dataset utilized is KDD1999 and might not detect new attacks |
RNN [51] | 94.12 | 98.19 | 95.13 | 93.56 | NA | When used with novel DDoS datasets, it might still show lower accuracy |
AIS-Based [52] | 96.56 | 95 | 96 | NA | NA | Although detection time is shorter, accuracy is also lower |
AE + DNN [53] | 98.43 | 99.22 | 97.12 | 98.57 | NA | It is a complex analysis, and when applied to fresh attacks, detection times may lengthen and accuracy may decline |
RDAER | 99.92 | 99.9 | 99.9 | 99.9 | 100 | Training and testing data can be further tuned to improve the accuracy |