| Mestry and Rathi | 2022 | Real-time malicious network detection in IoT | CNN-LSTM, CICFlowMeter | High detection accuracy | Needs feature extraction tools |
| Ismard | 2022 | Malicious network traffic detection | DL | Enhances security, reduces economic loss | No categorization or behavioral analysis |
| Islam et al. | 2022 | Framework for secure traffic classification | 1D-CNN, Flow-time-based features | High accuracy in encrypted traffic | – |
| Thirimanne et al.et al. | 2022 | Real-time IDS using DNN | DNN, NSL-KDD dataset | Effective feature extraction | Moderate accuracy |
| Gürfidan et al. | 2023 | ML/DL-based real-time anomaly detection | Blockchain + ML/DL | Enhances detection speed, security | Requires computational resources |
| Rohith Vallabhaneni Srinivas A Vaddadi | 2023 | CNN-RNN-based cyberattack detection | CNN, RNN | Captures spatial and temporal dependencies | High computational cost |
| Hattak et al. | 2023 | IoT intrusion detection using visualized network data | DL | Converts raw data into images | Lacks real-time analysis |
| Dabi Dabouabi Dalo Alionsi | 2023 | AI-driven real-time threat detection in IT networks | ML, DL | Effective for complex networks | Requires continuous updating |
| Liu et al | 2023 | Malicious traffic detection with FlowGAN | FlowGAN, DL | Enhances small sample detection | No threat categorization |
| Mei et al. | 2023 | DL-based anomaly detection | LetNet, LSTM | Robust, high-speed detection | Requires large datasets |
| Sharma et al. | 2023 | Autoencoder-based anomaly detection | Autoencoder | Learns complex patterns | No real-time classification |
| Alguliyev and Shikhaliyev | 2024 | Hybrid CNN-LSTM for network threat classification | CNN, LSTM | High classification accuracy | Requires large labeled datasets |
| Arjunan | 2024 | DL for anomaly detection in big data networks | CNN, LSTM, Transfer Learning | Handles large data volumes | Requires continuous training |
| Cadet et al. | 2024 | AI-powered surveillance threat detection | DL models | Applies to video feeds and sensors | No direct network traffic analysis |
| Faradias Izza Azzahra Faisal, et al. | 2024 | DL for OTT traffic classification | CNN, LSTM, Bi-LSTM | Effective QoS management | No threat categorization |
| Zhao et al. | 2024 | CNN-Focal-based IDS for real-time traffic detection | CNN-Focal | Addresses IDS limitations | Needs SoftMax tuning |
