Fit DKGE using feature-anchored subject kernels

Projects item-level kernels onto a shared anchor basis (via [dkge_build_anchor_kernels()]) and reuses [dkge_fit_from_kernels()] to enter the DKGE pipeline. Anchor provenance, coverage diagnostics, and subject item counts are stored in the resulting `dkge` object.

Usage

dkge_anchor_fit(
  features_list,
  K_item_list,
  folds = NULL,
  anchors = list(),
  design_kernel = NULL,
  dkge_args = list()
)

Arguments

features_list

List of subject feature matrices (`n_s x d` each).

K_item_list

List of subject item kernels (`n_s x n_s` each).

folds

Optional fold specification passed to [dkge_build_anchor_kernels()].

anchors

Named list overriding anchor-building defaults (`L`, `method`, `rho`, `fill`, `seed`, `sigma`, `center`, `whiten`, `eps`, `unit_trace`, `item_weights`).

design_kernel

Optional design kernel forwarded to [dkge_fit_from_kernels()]. Defaults to the identity.

dkge_args

Named list of additional arguments forwarded to [dkge_fit_from_kernels()].

Value

A `dkge` fit with anchor provenance under `fit$provenance$anchors`.

Examples

# \donttest{
set.seed(1)
features_list <- list(
  s1 = matrix(rnorm(30 * 5), 30, 5),
  s2 = matrix(rnorm(40 * 5), 40, 5),
  s3 = matrix(rnorm(35 * 5), 35, 5)
)
K_item_list <- lapply(features_list, function(X) {
  Z <- matrix(rnorm(nrow(X) * 4), nrow(X), 4)
  tcrossprod(Z)
})
fit <- dkge_anchor_fit(
  features_list, K_item_list,
  anchors = list(L = 8, method = "dkpp", seed = 1),
  dkge_args = list(rank = 2)
)
#> Warning: Subject 's1': beta matrix has reduced rank (4 < 8 effects).
#> Warning: Subject 's2': beta matrix has reduced rank (4 < 8 effects).
#> Warning: Subject 's3': beta matrix has reduced rank (4 < 8 effects).
dkge_anchor_diagnostics(fit)$summary
#> $method
#> [1] "dkpp"
#> 
#> $sigma
#> [1] 2.955892
#> 
#> $L
#> [1] 8
#> 
#> $mean_item_count
#> [1] 35
#> 
# }