In-sample yHat overstates accuracy. This masks a fold of phenotypes as NA, refits, and correlates held-out predictions against observed values.
Usage
kernel_cv(
y,
data,
random,
env,
gid,
folds = 5,
scheme = c("cv1", "cv0"),
seed = NULL,
tol = 1e-10,
scale_kernels = TRUE,
...
)Arguments
- y
character. Name of the response column in
data.- data
data.frame with the response, environment and genotype columns.
- random
named list of kernels passed to
kernel_model.- env
character. Name of the environment column in
data.- gid
character. Name of the genotype column in
data.- folds
integer. Number of cross-validation folds. Default 5.
- scheme
"cv1" (random cells) or "cv0" (leave-one-environment-out).
- seed
integer or NULL. Optional seed for reproducible fold assignment.
- tol
numeric. Eigenvalue tolerance passed to the kernel decomposition.
- scale_kernels
logical. If
TRUE(default) kernels are scaled before fitting.- ...
further arguments passed to
kernel_model.
Value
A list summarising out-of-sample predictive ability, with components
scheme (the validation scheme used), folds (number of folds),
accuracy (overall correlation between held-out predictions and
observations), rmse (overall root-mean-square error),
per_fold (a data.frame of per-fold accuracy and rmse),
fold_id (integer fold assignment for each row of data),
predicted (held-out predictions) and observed (the observed
response).