View source: R/cross_validate.R
set_cv | R Documentation |
locus
.The cross-validation procedure uses the variational lower bound as objective
function and is used to select the prior average number of predictors
p0_av
expected to be included in the model. p0_av
is used to
set the model hyperparameters and ensure sparse predictor selections.
set_cv(
n,
p,
n_folds,
size_p0_av_grid,
n_cpus,
tol_cv = 0.1,
maxit_cv = 1000,
verbose = TRUE
)
n |
Number of samples. |
p |
Number of candidate predictors. |
n_folds |
Number of number of folds. Large folds are not recommended for large datasets as the procedure may become computationally expensive. Must be greater than 2 and smaller than the number of samples. |
size_p0_av_grid |
Number of possible values of p0_av to be compared. Large numbers are not recommended for large datasets as the procedure may become computationally expensive. |
n_cpus |
Number of CPUs to be used for the cross-validation procedure. If large, one should ensure that enough RAM will be available for parallel execution. Set to 1 for serial execution. |
tol_cv |
Tolerance for the variational algorithm stopping criterion used within the cross-validation procedure. |
maxit_cv |
Maximum number of iterations allowed for the variational algorithm used within the cross-validation procedure. |
verbose |
If |
This cross-validation procedure is available only for
link = "identity"
.
An object of class "cv
" preparing the settings for the
cross-validation settings in a form that can be passed to the
locus
function.
locus
seed <- 123; set.seed(seed)
###################
## Simulate data ##
###################
## Example using small problem sizes:
##
n <- 150; p <- 200; p0 <- 50; d <- 25; d0 <- 20
## Candidate predictors (subject to selection)
##
# Here we simulate common genetic variants (but any type of candidate
# predictors can be supplied).
# 0 = homozygous, major allele, 1 = heterozygous, 2 = homozygous, minor allele
#
X_act <- matrix(rbinom(n * p0, size = 2, p = 0.25), nrow = n)
X_inact <- matrix(rbinom(n * (p - p0), size = 2, p = 0.25), nrow = n)
shuff_x_ind <- sample(p)
X <- cbind(X_act, X_inact)[, shuff_x_ind]
bool_x_act <- shuff_x_ind <= p0
pat_act <- beta <- matrix(0, nrow = p0, ncol = d0)
pat_act[sample(p0*d0, floor(p0*d0/5))] <- 1
beta[as.logical(pat_act)] <- rnorm(sum(pat_act))
## Gaussian responses
##
Y_act <- matrix(rnorm(n * d0, mean = X_act %*% beta, sd = 0.5), nrow = n)
Y_inact <- matrix(rnorm(n * (d - d0), sd = 0.5), nrow = n)
shuff_y_ind <- sample(d)
Y <- cbind(Y_act, Y_inact)[, shuff_y_ind]
########################
## Infer associations ##
########################
list_cv <- set_cv(n, p, n_folds = 3, size_p0_av_grid = 3, n_cpus = 1)
vb <- locus(Y = Y, X = X, p0_av = NULL, link = "identity", list_cv = list_cv,
user_seed = seed)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.