Table 1 Scenario 1: Single continuous incomplete variable missing at random, 9 continuous complete variables, 100 random runs, 30% missing rate
From: Secure distributed multiple imputation enables missing data inference for private data proprietors
Final analysis | Imputation | Performance | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
Technology | Θ bias | Θ SD | Θ rMSE | \(| \widehat{{\rm{y}}}-{\rm{y}}|\) (μ) | \(| \widehat{{\rm{y}}}-{\rm{y}}|\) (σ) | Discr. | \(| \widehat{{\rm{y}}}-{\rm{y}}|\) (μ) | \(| \widehat{{\rm{y}}}-{\rm{y}}|\) (σ) | Time (s) | Net (MB) | |
500 inds. | Python | 0.045 | 10-4 | 0.045 | 0.041 | 0.034 | 0 | 0.496 | 0.073 | 0.061 | N/A |
PyMICE | 0.042 | 1.57 × 10-16 | 0.042 | 0.040 | 0.032 | N/A | 0.498 | 0.034 | 0.223 | N/A | |
SMC-MICE | 0.045 | 4.18 × 10-4 | 0.045 | 0.041 | 0.034 | 0 | 0.496 | 0.073 | 0.090 | 4.737 | |
MHE-MICE | 0.065 | 3.91 × 10-4 | 0.065 | 0.044 | 0.035 | 0 | 0.496 | 0.071 | 1658 | 19,834 | |
5000 inds. | Python | 0.025 | 4.00 × 10-16 | 0.025 | 0.031 | 0.027 | 0 | 0.814 | 0.601 | 0.149 | N/A |
PyMICE | 0.019 | 4.15 × 10-16 | 0.019 | 0.027 | 0.025 | N/A | 0.813 | 0.600 | 0.205 | N/A | |
SMC-MICE | 0.019 | 1.15 × 10-4 | 0.019 | 0.029 | 0.026 | 0 | 0.813 | 0.601 | 0.536 | 44.641 | |
MHE-MICE | 0.019 | 1.13 × 10-4 | 0.019 | 0.029 | 0.026 | 0 | 0.813 | 0.601 | 2129 | 32,839 | |
