Table 4 The performance of IHVLR with different datasets
From: IHVFL: a privacy-enhanced intention-hiding vertical federated learning framework for medical data
Related works | Scheme | Without coordinator | Diabetes dataset | Breast cancer dataset | ||||
|---|---|---|---|---|---|---|---|---|
Accuracy (%) | AUC | Runtime (s) | Accuracy (%) | AUC | Runtime (s) | |||
Baseline | Plaintext | – | 78.355 | 0.864 | 0.095 | 98.246 | 0.999 | 0.069 |
HECLR (Hardy et al 2017) | HE | ✗ | 77.922 | 0.864 | 75.552 | 96.491 | 0.999 | 107.251 |
HELR (Yang et al 2019b) | HE | ✓ | 78.355 | 0.865 | 33.273 | 97.661 | 0.999 | 25.265 |
SSHELR (Chen et al 2021) | SS+HE | ✓ | 78.355 | 0.864 | 93.014 | 97.076 | 0.999 | 153.946 |
Ours | SS+HE | ✓ | 77.922 | 0.864 | 190.423 | 97.076 | 0.999 | 265.623 |