Table 2.
Effect of percentage of missing data on the MRdAE (median relative absolute error) for the four methods applied to the two subdatabases. For each method, a one-way Kruskal–Wallis test was conducted to test the effect of the percentage of missing data on the MRdAE. The P values are presented in the table for each method and each trait
| Traits |
|||||||||
|---|---|---|---|---|---|---|---|---|---|
| Methods | H | LDMC | LM | LS | RH | SM | SNP | SS | SLA |
| Whole subdatabase | |||||||||
| Average | 0.55 | 0.44 | 0.24 | 0.11 | 0.34 | 0.38 | 0.06 | 0.48 | 0.38 |
| Median | 0.72 | 0.22 | 0.42 | 0.97 | 0.33 | 0.55 | 0.01 | 0.66 | 0.37 |
| MICE | 0.22 | 0.00 | 0.00 | 0.00 | 0.11 | 0.00 | 0.53 | 0.01 | 0.00 |
| Dissimilarity | 0.46 | 0.87 | 0.00 | 0.11 | 0.25 | 0.00 | 0.01 | 0.01 | 0.00 |
| Relationships | 0.00 | 0.00 | 0.00 | 0.00 | 0.02 | 0.00 | 0.40 | 0.00 | 0.00 |
| Herbaceous subdatabase | |||||||||
| Average | 0.69 | 0.55 | 0.07 | 0.07 | 0.47 | 0.32 | 0.01 | 0.80 | 0.95 |
| Median | 0.89 | 0.03 | 0.00 | 0.19 | 0.20 | 0.01 | 0.00 | 0.46 | 0.89 |
| MICE | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.36 | 1.00 | 0.01 | 0.00 |
| Dissimilarity | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.52 | 0.00 | 0.00 | 0.07 |
| Relationships | 0.00 | 0.31 | 0.00 | 0.00 | 0.00 | 0.00 | 0.36 | 0.39 | 0.16 |
H, vegetative height; LDMC, leaf dry matter content; LM, leaf mass; LS, leaf surface; MICE, Multivariate Imputation by Chained Equations; RH, reproductive height; SM, seed mass; SNP, seeds number per plant; SS, seed shape; SLA, specific leaf area.
Significant P values (P < 0.05) are in bold.