Table 3.
Maximum accuracies obtained among four training set optimization methods in predictions made for each combination of trait and family
| Family | Trait | Accuracy | TS size | Method |
|---|---|---|---|---|
| FjDe | ALD | 0.23 | 77 | Mean relationship |
| FNP | 0.18 | 21 | Max relationship | |
| PC1 | 0.26 | 77 | Mean relationship | |
| PC2 | 0.36 | 56 | Max relationship | |
| FjPi | ALD | 0.36 | 189 | Mean relationship |
| FNP | 0.59 | 174 | Max relationship | |
| PC1 | 0.36 | 178 | Max relationship | |
| PC2 | 0.26 | 202 | Mean relationship | |
| FjPL | ALD | 0.10 | 130 | CDmean-opt |
| FNP | 0.16 | 22 | Max relationship | |
| PC1 | 0.20 | 120 | Max relationship | |
| PC2 | 0.22 | 10 | CDmean-opt | |
| GaPi | ALD | 0.46 | 156 | Mean relationship |
| FNP | 0.40 | 13 | Max relationship | |
| PC1 | 0.54 | 191 | Clusters | |
| PC2 | 0.28 | 19 | Mean relationship | |
| GaPL | ALD | 0.72 | 136 | Clusters |
| FNP | 0.78 | 37 | Max relationship | |
| PC1 | 0.81 | 129 | Mean relationship | |
| PC2 | 0.40 | 140 | Max relationship | |
| GDFj | ALD | 0.21 | 66 | Mean relationship |
| FNP | 0.32 | 15 | Max relationship | |
| PC1 | 0.19 | 31 | Mean relationship | |
| PC2 | 0.01 | 10 | Mean relationship |
TS training set, ALD acoustic linear distance, ANP number of acoustic peaks, FNP number of force peaks, PC1 principal component 1 (synthetic trait), PC2 principal component 2 (synthetic trait)