Skip to main content
. Author manuscript; available in PMC: 2019 Jun 19.
Published in final edited form as: J Am Stat Assoc. 2018 Jun 19;113(524):1710–1721. doi: 10.1080/01621459.2017.1361830

Figure 2:

Figure 2:

Comparison of the mean square prediction error (MSPEs) for Bayes Ridge (BRR), Bayes Lasso (BL), Bayes LMM (BLMM), Support Vector Machine (SVM), and BAKR approximating Gaussian kernels with bandwidth parameter values h = {5, 2,1, 0.5, 0.01}. In Scenario I, pairwise interactions make up 80% of the broad-sense heritability (i.e. ρ = 0.2). In Scenario II, additive effects dominate 80% of the broad-sense heritability (i.e. ρ = 0.8). These results are based on 100 different simulated datasets in both scenarios.