Prints a formatted evaluation table including per-trait metrics,
gate calibration, and conformal coverage. For a trait represented by
multiple latent columns, the displayed gate uses its first latent column.
Requires the original
pigauto_data object to compute test-set performance.
Usage
# S3 method for class 'pigauto_fit'
summary(object, ..., data = NULL)Value
Invisibly returns the evaluation data.frame, or NULL if
data is not supplied, the fit has no test cells, or evaluation
fails. If evaluation fails, the method prints a message and returns
NULL.
Examples
# \donttest{
data(avonet300, tree300)
tree <- ape::keep.tip(tree300, tree300$tip.label[seq_len(30L)])
traits <- avonet300[match(tree$tip.label, avonet300$Species_Key),
c("Mass", "Wing.Length"), drop = FALSE]
rownames(traits) <- tree$tip.label
pd <- preprocess_traits(traits, tree)
splits <- make_missing_splits(pd$X_scaled, trait_map = pd$trait_map)
fit <- fit_pigauto(pd, tree, splits = splits, epochs = 5L,
verbose = FALSE)
#> Warning: Small validation set for 2 trait(s): Mass (n=2), Wing.Length (n=1). Calibrated gate and conformal scores will be noisy for these trait(s). 95% split-conformal coverage is NOT achievable for 2 trait(s) with fewer than 19 validation cells (Mass (n=2), Wing.Length (n=1)): the achievable ceiling is n_val / (n_val + 1), which only reaches 0.95 at n_val >= 19. See `?fit_pigauto` under 'Calibration at small n' for smoothing options.
# This small test set illustrates the output; it is too small for a
# stable performance conclusion. Use a held-out design sized for your
# question. Repeated splits or seeds assess split/training variability,
# but do not by themselves establish generalisation.
summary(fit, data = pd)
#> pigauto_fit summary
#> --------------------------------------------------------
#> Species: 30 | Traits: 2 | Epochs: 0
#> GNN: off (baseline only)
#> Trait types: continuous=2
#>
#> Best validation loss: 0.1865
#> Test loss: 0.1262
#>
#> Trait Performance (test set):
#> --------------------------------------------------------
#> Trait Type n RMSE r MAE
#> -------------------------------------------------------------------------
#> Mass continuous 5 0.072 0.998 0.050
#> Wing.Length continuous 7 0.154 0.964 0.142
#>
#> Baseline (BM) for comparison:
#> -------------------------------------------------------------------------
#> Mass continuous 5 0.072 0.998 0.050
#> Wing.Length continuous 7 0.154 0.964 0.142
#>
#> Gate Calibration:
#> Mass: 0.000 Wing.Length: 0.000
#>
#> Conformal Scores (latent scale, 95% nominal):
#> Mass: 0.3205 Wing.Length: 0.0099
#> Conformal Coverage (test set):
#> Mass: 100.0% Wing.Length: 0.0%
#>
# }