Table 1 For each of the data sets, the mean relative errors (%) from each of the genomic prediction methods conducted using four random cross-validation splits with 90% training data and 10% validation data. The number of feature selected markers for VBS-ML is given in parentheses
From: Improved genomic prediction using machine learning with Variational Bayesian sparsity
Year | Methods | 1 | 2 | 3 | 4 | Ave. |
|---|---|---|---|---|---|---|
2014 | LMM | 7.38 | 7.21 | 7.44 | 7.31 | 7.34 |
BayesA | 7.28 | 7.16 | 7.37 | 7.25 | 7.27 | |
BayesB | 7.20 | 7.09 | 7.36 | 7.21 | 7.22 | |
Naive-ML | 7.61 | 7.56 | 8.31 | 7.61 | 7.77 | |
VBS-ML (354) | 6.49 | 7.08 | 7.58 | 7.04 | 7.05 | |
2016 | LMM | 5.36 | 5.00 | 5.11 | 5.23 | 5.18 |
BayesA | 5.36 | 5.02 | 5.15 | 5.20 | 5.18 | |
BayesB | 5.30 | 5.02 | 5.07 | 5.28 | 5.16 | |
Naive-ML | 6.67 | 6.28 | 6.59 | 6.10 | 6.41 | |
VBS-ML (409) | 5.20 | 4.89 | 4.94 | 5.04 | 5.02 | |
2017 | LMM | 3.53 | 3.37 | 3.63 | 3.45 | 3.50 |
BayesA | 3.54 | 3.38 | 3.64 | 3.48 | 3.51 | |
BayesB | 3.51 | 3.43 | 3.65 | 3.43 | 3.51 | |
Naive-ML | 3.99 | 4.16 | 4.18 | 3.77 | 4.03 | |
VBS-ML (315) | 3.48 | 3.26 | 3.54 | 3.26 | 3.39 | |
2018 | LMM | 5.94 | 4.98 | 5.80 | 5.94 | 5.67 |
BayesA | 5.94 | 4.99 | 5.77 | 5.94 | 5.66 | |
BayesB | 5.98 | 5.06 | 5.63 | 5.97 | 5.66 | |
Naive-ML | 6.39 | 5.70 | 6.20 | 6.23 | 6.13 | |
VBS-ML (385) | 5.89 | 4.98 | 5.16 | 5.86 | 5.47 |