Table 1 Summary of existing methods.
Method | Reference | Privacy preservation approach | Outcomes |
|---|---|---|---|
Expert system using federated learning | |||
MABC-SVM for HSP | Decentralized aggregation | Optimal feature selection and classification of heart disease | |
Federated Matched Averaging for HSP | Federated learning to enhance data privacy | Detect COVID-19 from a single chest X-ray image within seconds, while ensuring data privacy | |
FLIDS-BSAFSC | Decentralized training to reduce privacy risks | Classify, detect, and defend against attacks in IoT datasets | |
Decentralized FL (DFL) | Decentralized model aggregation | Minimizes dependency on a central entity, allowing flexible training across diverse federations of devices | |
Privacy preservation based expert system | |||
Privacy-preservation signaling game | Signaling Q-learning algorithm to secure data | Achieves convergent equilibrium and practical game parameters, protecting data in edge-computing-based IoT networks | |
ISD-k-ADP | Sensitivity Drift-based k-Anonymized Data Perturbation Scheme | Facilitates hiding EHR data with controlled noise, enabling effective and efficient classification through Two Stage Bagging Pruning based Ensemble | |
Blockchain-based Authentication | Blockchain for identity storage and three-factor authentication with Chebyshev chaotic map | Ensures secure user login and authentication | |
Improved Matrix Factorization (IMFPM) | Piecewise Mechanism (PM) with random projection technology | Protects privacy of rating values and item sets, while reducing the influence of privacy noise on estimation error | |
