Table 1 Features and challenges.

From: Enhanced IoT threat detection using Graph-Regularized neural networks optimized by Sea-Lion algorithm

Author [citation]

Features

Challenges

Farhan Ulla et al.16

Lower access latency

High precision but Lower F-measure

Costly framework.

Has to provide security against various identified attacks

Keping Yu et al.17

Safe from the multiple malicious attacks

Low transmission cost, transmission delay, and signaling overhead

Low accuracy

High energy dissipated

Mohamed Amine Ferrag et al.18

Improve the delay response

Higher precision but relatively low F-measure.

Edge-IIoTset dataset

Incurs extra computational costs of signing and verifying the

Signatures

Yakub Kayode Saheed et al.19

Resistant against DoS attack and man-in-the-middle.

High RoC Curve

Feature Extraction

Eirini Anthi et al.20

Jacobian-based Saliency Map attack

ICS cybersecurity

Improved sensitivity

Reduces the life span of the network