Compute Subject-Specific Parameters

Usage

.tmb_compute_subject_pars(
  coefficients,
  u_hat,
  subject_levels,
  re_parsed,
  has_k,
  equation,
  price,
  subject_id,
  k_fixed = NULL,
  X_q0 = NULL,
  X_alpha = NULL,
  Z_q0 = NULL,
  Z_alpha = NULL,
  validate_subject_pars = TRUE
)

Arguments

coefficients: Named coefficient vector.
u_hat: Random effects matrix; columns ordered [block1_q0, block1_alpha, block2_q0, ...].
subject_levels: Character vector of subject IDs.
re_parsed: Canonical RE block structure.
has_k: Logical.
equation: Character.
price: Numeric vector of prices.
subject_id: Integer vector of 0-indexed subject IDs.
k_fixed: Numeric or NULL.
X_q0, X_alpha: Fixed-effect design matrices.
Z_q0, Z_alpha: Random-effect design matrices.
validate_subject_pars: Logical (default TRUE); see the validate_subject_pars argument of fit_demand_tmb() for the NA fallback.

Value

Data frame of subject-specific parameters.

Note

For factor-expanded RE specs (e.g. pdDiag(Q0+alpha~condition)), subject-level Q0 / alpha here use the first observed row of Z_q0 / Z_alpha per subject – which encodes the subject's first observed condition only. This helper does not emit per-(subject, condition) rows; use predict() for cell-level values.