Extract Random Effects from TMB Model

Returns subject-level random effect deviations on the natural (log) scale. These are the Cholesky-transformed deviations, not standardized scores. To obtain the standardized random effects (u matrix), access object$tmb_obj directly.

Usage

# S3 method for class 'beezdemand_tmb'
ranef(object, ...)

Arguments

object

A beezdemand_tmb object.

...

Additional arguments.

Value

Data frame with subject-level random effects. Columns:

• id: subject identifier

• b_i, c_i (when present): first-column convenience aliases for q0_(Intercept) and alpha_(Intercept). Preserved for backward compatibility with older callers.

• q0_<term>: per-block random-effect coefficients for log-Q0, one column per random-effects design column from the parsed block structure. For factor-expanded or multi-block fits, these expose the per-condition slope REs that b_i / c_i alone do not surface.

• alpha_<term>: analogous columns for log-alpha.

Examples

# \donttest{
data(apt)
fit <- fit_demand_tmb(apt, equation = "exponential", verbose = 0)
#>   equation='exponential': Dropped 14 zero-consumption observations (146 remaining).
head(nlme::ranef(fit))
#>    id        b_i        c_i q0_(Intercept) alpha_(Intercept)
#> 1  19  0.4351751 -0.6116759      0.4351751        -0.6116759
#> 2  30 -0.8297240  0.4438124     -0.8297240         0.4438124
#> 3  38 -0.3632195  0.1170514     -0.3632195         0.1170514
#> 4  60  0.3962798  0.1156735      0.3962798         0.1156735
#> 5  68  0.4495371 -0.3215051      0.4495371        -0.3215051
#> 6 106 -0.1525620  0.4772347     -0.1525620         0.4772347
# }