Table 2 Comparison of different algorithms on synthetic data, NMSE is scaled by \(10^{-2}\), CPU time is in seconds, and “sr” denotes the sample ratio
m = 1000 sr = 6.91% | m = 2000 sr = 3.80% | m = 3000 sr = 2.67% | m = 5000 sr = 1.70% | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
NMSE | Rank | Time | NMSE | Rank | Time | NMSE | Rank | Time | NMSE | Rank | Time | |
APGL | 3.82 | 68 | 45.18 | 3.81 | 104 | 447.22 | 3.66 | 132 | 2372.02 | 3.49 | 179 | 21448.57 |
AIS-impute | 3.05 | 5 | 5.09 | 3.61 | 80 | 259.82 | 3.53 | 80 | 500.96 | 3.33 | 80 | 1579.33 |
ASD | 4.75 | 5 | 1.60 | 4.81 | 5 | 3.29 | 4.73 | 5 | 5.07 | 4.77 | 5 | 9.37 |
IRNN-TNN | 1.83 | 5 | 92.85 | 1.77 | 5 | 1046.51 | 1.70 | 5 | 4761.42 | 1.67 | 5 | 23906.09 |
FaNCL-LSP | 1.86 | 5 | 2.15 | 1.78 | 5 | 8.81 | 1.73 | 5 | 19.12 | 1.61 | 5 | 39.94 |
DNNR(\(p=2/3\)) | 1.84 | 5 | 53.04 | 1.77 | 5 | 511.33 | 1.72 | 5 | 3115.31 | 1.63 | 5 | 37198.66 |
DLRL | 1.85 | 5 | 32.87 | 1.78 | 5 | 170.94 | 1.71 | 5 | 627.53 | 1.61 | 5 | 15043.28 |
Proposed | 1.83 | 5 | 0.36 | 1.76 | 5 | 1.3467 | 1.70 | 5 | 2.0432 | 1.61 | 5 | 5.13 |