Table 2.
ML-based privacy works in IoT environments—Centralized Learning.
| Reference | Attack | Privacy Protection Model |
ML Model |
Dataset | Results | Scenario |
|---|---|---|---|---|---|---|
| Rahulamathavan [54] (2014) |
Data leakage | Encrytion (Classification) |
SVM | WBC PID IRIS JAFFE |
ACC 98.24% 86.98% 87.33% 89.67% |
Cloud Computing |
| Wang [57] (2017) |
Data leakage |
Encrytion (Classification) |
ML-ELM | MNIST | ACC 79.83% (AES) 90.44%(DES) |
Cloud Computing |
| Zhu [59] (2017) |
Data leakage | Encrytion (Classification) |
SVM | PID | ACC 94% |
IoT eHealth |
| Jiang [61] (2019) |
Data leakage | Encrytion (Training) |
Homomorphic surf and fast image matching |
DR1 RetiDB Messidor |
AUC [86%, 89%] |
IoT eHealth |