Table 3.

Deep learning models for IoT applications.

Ref.	Approach	Description	Results
[47]	TensorFlow DNN	Identify pirated software through source code copying	97.46% classification accuracy
[48]	Deep learning framework for secure smart city	Utilize blockchain for decentralized communication in CPS	Precision: 0.7244, Recall: 0.7078, F1 score: 0.7118
[49]	FFDNN Wireless IDS with WFEU	Intrusion detection system equipped with Wireless Feature Extraction Unit	Binary classification: 87.10% accuracy, multiclass classification: 77.16% accuracy
[50]	Sequential methodology with Text-CNN and GRU	Collect network layer and application layer attributes for intrusion detection	F1 score: 0.98
[51]	Deep learning-based IDS for IoT networks	Categorize data flow for multiclass and binary classification	NSL-KSS dataset: 99.5% accuracy, CIDDS-001 dataset: 99.3% accuracy, UNSWNB15 dataset: 99.1% accuracy
[52]	IoT-IDCS-CNN	Harness convolutional neural networks for intrusion detection	Binary classification: 99.30% accuracy, multiclass classification: 98.20% accuracy