Advances, Systems and Applications
From: Intrusion detection systems for IoT-based smart environments: a survey
Technique | Advantages | Disadvantages |
---|---|---|
Data mining | 1- Models are created automatically | 1- Based on historical data |
 | 2- Applicable in different environments | 2- Depends on complex algorithms |
 | 3- Suitable for online datasets |  |
Machine learning | 1- High detection accuracy | 1- Requires training data |
 | 2- Suitable for massive data volumes | 2- Long training time |
Statistical model | 1- Suitable for online datasets | 1- Based on historical behavior |
 | 2- System simplicity | 2- Detection accuracy depends on statistical and mathematical operations |
Rule model | 1- Suitable for online datasets | 1- Based on a set of rules |
 | 2- System simplicity | 2- High false positive rate |
Payload model | 1- High detection accuracy for known attacks | 1- Privacy issues |
 |  | 2- Long processing time |
Protocol model | 1- High detection accuracy for a specific type of attack | 1- Designed for a specific type of protocol |
Signal processing model | 1- High detection accuracy | 1- Depends on complex pattern-recognition methods |
 | 2- Low false positive rate |  |