Table 8 Comparison of proposed Hybrid C4.5–DQN IDS with selected recent IDS approaches.
Model | Dataset | Accuracy (%) | Precision (%) | Recall (%) | F1-Score (%) | Remarks |
|---|---|---|---|---|---|---|
HIDS-RPL86 | CIC-DDoS2019 | 99.87 | 98.50 | 98.64 | 98.54 | Hybrid CNN + LSTM for RPL-based IoMT; optimized for low-power environments |
Meta-Learning Ensemble IDS87 | IoMT-specific datasets (various feature sizes) | 99.50 | 99.40 | 99.60 | 99.50 | Model dynamically reweights classifiers |
PSO-AdaBoost IDS88 | NSL-KDD | – | – | 96.67 | – | Particle Swarm Optimization for feature selection and AdaBoost for classification (Accuracy/Precision values not numerically reported) |
HIDS (GA-DT)89 | NSL-KDD | 99.88 | – | – | – | Genetic Algorithm (GA) – Decision Tree (DT) model |
StandAlone DQN | CICIoMT-2024 | 99.70 | 98.31 | 98.81 | 98.60 | Dynamic Learning ability of DQN |
Hybrid C4.5–DQN (Proposed) | CICIoMT-2024 | 99.40 | 99.65 | 99.65 | 99.48 | Combines the speed/efficiency of C4.5 with the adaptive learning of DQN |
