Table 2.

Significance of machine learning in cyber-security.

Techniques	Problems	Findings	Advantages	References
IF and convolutional neural network (CNN) models	Hybrid cyber-attack detections	Detected the maximum attacks with maximum accuracy	Proposed hybrid model of reasonable efficiency but lacks comparison	[64]
NB, RF, and J48 model	To detect DDoS attacks	Random forest model is more accurate than other models	Efficient attack detection in SCADA system but limited to one attack type	[65]
DT model	User-centric security and fake data identification for IoT-based critical infrastructure	To find theoretical and experimental solutions that solve security issues	Secure channel by decision tree in IoT security. Lacks comparative study	[66]
RF, DT, SVM, perceptron, and Gaussian NB classifier	Secure data dissemination architecture	Accurate, secure, and reliable architecture for IoT-based critical infrastructure	Efficient cyber-security in critical infrastructure but exhibited less accuracy	[67]
RF, SVM, MLP, AdaBoost and hybrid model	Cyber-threat detection from real-time dataset	Efficient threat hunting with high accuracy and precision; AdaBoost outperformed all models	All models performed well for cyber-threat detection. Comparative analysis shows high effectiveness of AdaBoost, RF, and hybrid models for real-world application.	This study