Predict from a TMB mixed-effects discounting model

Usage

# S3 method for class 'beezdiscounting_tmb'
predict(
  object,
  newdata = NULL,
  type = c("response", "parameters"),
  level = "subject",
  ...
)

Arguments

object: A beezdiscounting_tmb object.
newdata: Optional data frame. NULL uses the fitting data. A supplied newdata must use the package's canonical column names (id, x, y), regardless of the id_var/x_var/y_var you gave at fit time.
type: "response" (fitted indifference proportions on (0,1)) or "parameters" (the per-subject parameter tibble).
level: For type = "response": "subject" (default; conditions on each subject's estimated random intercept, requires the id column) and/or "population" (random effects set to zero - the population-mean curve; no id column needed). Pass c("population", "subject") for both columns side-by-side. A numeric nlme-style level is rejected with an error. For an s-target 2-RE fit (k + s ~ 1), the subject level uses each subject's estimated s_i; the population level (random effects zero) uses the soft-clamped population s = .dd_soft_clamp_s_log(log_s), matching the kernel near the (0.05, 20) bounds (equal to exp(log_s) in the interior).
...: Unused.

Value

type = "parameters": the subject-parameter tibble (one row per subject).
type = "response", level = "subject" only: newdata as a tibble plus a .fitted column.
type = "response", "population" requested: newdata plus predict.fixed (and predict.id when "subject" is also requested), matching the nlme::predict.lme(level = 0:1) column-name convention so nlme-based plotting code runs unchanged.

Examples

# \donttest{
dd <- simulate_dd_ip(n_subjects = 20, seed = 1)
fit <- fit_dd_tmb(dd, equation = "mazur", family = "sltb",
                  random_effects = k ~ 1, verbose = 0)

# Subject-conditional fitted values
head(predict(fit, type = "response"))
#> # A tibble: 6 × 4
#>   id        x     y .fitted
#>   <chr> <dbl> <dbl>   <dbl>
#> 1 1         7 0.998  0.953 
#> 2 1        30 0.891  0.827 
#> 3 1       180 0.570  0.443 
#> 4 1       365 0.290  0.282 
#> 5 1       730 0.152  0.164 
#> 6 1      1460 0.145  0.0892

# Population-mean curve at specific delays (no id column needed)
nd <- data.frame(x = c(7, 30, 180, 365))
predict(fit, newdata = nd, level = "population")
#> # A tibble: 4 × 2
#>       x predict.fixed
#>   <dbl>         <dbl>
#> 1     7         0.930
#> 2    30         0.756
#> 3   180         0.341
#> 4   365         0.203

# Per-subject parameters
head(predict(fit, type = "parameters"))
#> # A tibble: 6 × 3
#>   id        u_i       k
#>   <chr>   <dbl>   <dbl>
#> 1 1     -0.811  0.00699
#> 2 10    -0.709  0.00738
#> 3 11     1.33   0.0218 
#> 4 12     0.0467 0.0110 
#> 5 13    -0.543  0.00806
#> 6 14    -2.30   0.00317
# }