This exists for the tests: the MME identity against the V form, and the score against central finite differences. It is not the user-facing interface.
Usage
eval_internal(
formula,
data,
pedigree = NULL,
theta,
missing_code = NULL,
with_dense = TRUE,
metafounders = NULL,
gamma = NULL,
weights = NULL,
genotypes = NULL,
blend = 0.05,
apy_core = NULL,
vecchia_k = NULL
)Arguments
- formula
the same as in model()
- data
data.frame
- pedigree
data.frame or NULL
- theta
vector of components at which to evaluate
- missing_code
missing-value code or NULL
- with_dense
TRUE also computes the dense V form, which only handles a small problem; there is no size limit here, unlike
eval_internal_ar1(), whose default skips it above 2000 cells- metafounders
as in
model()- gamma
as in
model()- weights
as in
model()- genotypes
as in
model(): the single-step path is available here too, so a likelihood loop written outside the package can walk the genomic model- blend
as in
model()- apy_core
as in
model()- vecchia_k
as in
model()