Table 9.
Relationship between outbreak issues and danger theory as well as the benefits of the outbreak detection model based on danger theory.
| Outbreak issues | Effect | Danger theory characteristics | Outbreak detection model based on danger theory | Benefits |
|---|---|---|---|---|
| Outbreak and outbreak period is hard to determine. Most of the time, only outbreaks period are available for model development | Models lose robustness when unseen outbreak patterns vary from those of the trained model. This reduces detection capability when the unseen patterns vary from those in the trained model | Antigens have a similar chance of becoming infected with a harmful pathogen | No need to define outbreak and non-outbreak class. In other words, no training phase involved | Improves detection performance by balancing the DR and FAR Offers a new method for multivariate detection Has potential to be applied as a real-time system as applied in intrusion detection areas |
| Weak and inconsistent outbreak signal | Imbalance in the result between DR and FAR | The dendrite cell process uses multiple input signals | Accepts multiple input factors to improve detection performance rather than relying on a single predictive factor |