tidy.ref_fit

Usage

tidy.ref_fit(x, ...)

glance.ref_fit(x, ...)

glance.ref_assessment(x, ...)

augment.ref_fit(x, newdata, uncertainty = c("total", "conditional"), ...)

tidy() gives one row per outcome: family, fit status, sample size, total effective degrees of freedom and one edf_<term> column per smooth term, and any engine message. glance() gives one row per fit (or per assessment, averaging the overall and marginal tables over outcomes). augment() returns newdata with .z_<outcome> and .centile_<outcome> columns and the covariate .support; the result can be passed straight to ref_joint().

Arguments

Argument	Description
x	A ref_fit or ref_assessment.
...	Passed to predict.ref_fit() by augment(); unused otherwise.
newdata	Data frame of target observations.
uncertainty	Passed to predict.ref_fit().

Value

A tibble.

Examples

ref <- ref_simulate(120, seed = 3)
fit <- ref_fit(ref_spec(ref_gaussian(), ~ s(age, k = 5) + sex), ref, "y")
tidy(fit)

# A tibble: 1 × 7
  outcome family   status     n   edf `edf_s(age)` message
  <chr>   <chr>    <chr>  <int> <dbl>        <dbl> <chr>
1 y       gaussian ok       120  3.64         1.64 <NA>

glance(fit)

# A tibble: 1 × 8
  family   engine     n n_outcomes  n_ok covariates adapted calibrated
  <chr>    <chr>  <int>      <int> <int> <chr>      <lgl>   <lgl>
1 gaussian mgcv     120          1     1 age, sex   FALSE   FALSE

augment(fit, ref[1:3, ], uncertainty = "conditional")

# A tibble: 3 × 10
    age sex   site      y marker_01 marker_02 marker_03    .z_y .centile_y
  <dbl> <fct> <fct> <dbl>     <dbl>     <dbl>     <dbl>   <dbl>      <dbl>
1  30.1 M     B      8.83      9.45     11.2       9.83  0.0867      0.535
2  68.5 F     A      9.98      9.72      9.07     10.1  -0.854       0.196
3  43.1 F     A      9.91      9.82     10.6      11.3   0.615       0.731
# ℹ 1 more variable: .support <chr>