Balancing Data Privacy and 5G VNFs Security Monitoring: Federated Learning with CNN + BiLSTM + LSTM Model

Table 12

Comparison with previous studies that used FL for intrusion detection.

Reference	Year	Technique used	Dataset utilized	Accuracy (%)	Precision (%)	Encryption technique
[36]	2022	Multi-layer perceptron (MLP)	Private dataset	95	—	None
[37]	2020	Transfer learning	CICIDS2017	91.93	—	None
[39]	2022	ANN based-FL	UNSW-NB15	93.6	—	None
[40]	2022	Neural network that uses transformer	NSL-KDD	99.48	99.49	None
[41]	2023	Gower dissimilarity matrices	TON_IOT	95.5	96	None
[54]	2020	MLP	NSL-KDD	98.11	—	None
[55]	2021	CNN + GRU	Industrial CPS	99.20	98.86	None
Our model		CNN + BiLSTM + LSTM	InSDN; CICIDS2017	99.98; 99.58	99.98; 99.38	TLS V 1.3