Table 1 Various machine learning methods for IoT systems.
From: Improved model for intrusion detection in the Internet of Things
ML Method | Attacks | Pros | Cons |
|---|---|---|---|
U2R, R2L, Flooding attacks, DoS, DDoS | Simplicity | There are two challenges: one is to find the best value of K and identify missing nodes | |
Easy and simple to use method | It needs ample storage Ease to use if DTs are few, but it’s always complex in computation | ||
Scan, DDoS (TCP, UDP flood), smurf, port sweep | Data with several features SVMs use less storage and memory | It isn’t easy to understand and interpret |
