Default hyperparameters of the Bayesian ridge regression prior

Formula: S = Var(Y)(1 - R2)(nu + 2); S_beta = Var(Y) R2 (nu_beta + 2)

Usage

Brrhyper(y, R2 = 0.5, nu = 5, nu_beta = 5)

Arguments

y: Response vector of length n.
R2: Prior proportion of the phenotypic variance explained by the markers.
nu: Degrees of freedom of the scaled inverse chi-square prior on the residual variance.
nu_beta: Degrees of freedom of the prior on the marker-effect variance.

Value

List with S, S_beta, nu, nu_beta, var_y, n.

References

Montesinos Lopez, Montesinos Lopez and Crossa (2022), Multivariate Statistical Machine Learning Methods for Genomic Prediction, Springer, doi:10.1007/978-3-030-89010-0. Chapter 6, the BGLR default rules quoted on pp. 175 and 184: the scale of the residual prior is Var(Y)(1 - R2)(nu + 2) and, for the BRR, the scale of the marker-effect prior is Var(Y) R2 (nu_beta + 2).

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
Brrhyper(V)
#> $S
#> [1] 21
#> 
#> $S_beta
#> [1] 21
#> 
#> $nu
#> [1] 5
#> 
#> $nu_beta
#> [1] 5
#> 
#> $var_y
#> [1] 6
#> 
#> $n
#> [1] 8
#> 
#> $method
#> [1] "BRR prior hyperparameters, MVSML Chap. 6 pp. 175, 184"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"