Skip to contents

Models

pptr()
Trains a Projection-Pursuit oblique decision tree.
pprf()
Trains a Random Forest of Projection-Pursuit oblique decision trees.

Prediction and fitted values

predict(<pprf_classification>)
Predicts labels or vote proportions from a pprf model (classification mode).
predict(<pprf_regression>)
Predicts numeric responses from a pprf model (regression mode).
predict(<pptr_classification>)
Predicts labels or per-group one-hot proportions from a pptr model (classification mode).
predict(<pptr_regression>)
Predicts numeric responses from a pptr model (regression mode).
fitted(<ppmodel>)
Fitted (in-sample) predictions from a ppforest2 model.
residuals(<ppmodel>)
Residuals from a regression ppforest2 model.
nobs(<ppmodel>)
Number of observations used to fit a ppforest2 model.
formula(<ppmodel>)
Formula extractor for ppforest2 models.

Printing and summaries

print(<pptr>)
Prints the structure of a pptr tree.
print(<pprf>)
Prints a compact summary of a pprf forest.
summary(<pprf>)
Summary of a pprf forest (shared header + VI).

Out-of-bag results

oob_error()
Out-of-bag error for a random forest.
oob_predictions()
Out-of-bag predictions for a random forest.
oob_samples()
Out-of-bag row indices per tree.
bag_samples()
In-bag row indices per tree.

Variable importance

projection_importance()
Projection-coefficient variable importance.
weighted_importance()
Weighted projection variable importance for a random forest.
permuted_importance()
Permuted variable importance for a random forest.

Plotting

plot(<pptr>)
Plot a pptr model.
plot(<pprf>)
Plot a pprf model.

Training strategies

Each step of node training is a strategy that can be replaced. See vignette("custom-strategies").

pp_pda()
PDA projection pursuit strategy.
vars_all()
All-variables selection strategy.
vars_uniform()
Uniform random variable selection strategy.
cutpoint_mean_of_means()
Mean-of-means split cutpoint strategy.
stop_any()
Composite stopping rule (logical OR).
stop_max_depth()
Maximum-depth stopping rule.
stop_min_size()
Minimum-size stopping rule.
stop_min_variance()
Minimum-variance stopping rule.
stop_pure_node()
Pure-node stopping rule.
binarize_disabled()
Disabled binarization strategy (placeholder).
binarize_largest_gap()
Largest-gap binarization strategy.
grouping_by_cutpoint()
Cutpoint-based grouping strategy (regression).
grouping_by_label()
Label-based grouping strategy.
leaf_majority_vote()
Majority-vote leaf strategy.
leaf_mean_response()
Mean-response leaf strategy.

tidymodels integration

pp_tree()
Parsnip model specification for pptr.
pp_rand_forest()
Parsnip model specification for pprf.
update(<pp_tree>)
Update a pp_tree model specification.
update(<pp_rand_forest>)
Update a pp_rand_forest model specification.

Saving and loading models

save_json()
Save a model to a JSON file.
load_json()
Load a model from a JSON file.

Datasets

california_housing
California Housing Dataset
crab
Australian Crabs Dataset
crabs
Crabs Dataset
fishcatch
Fish Catch Dataset
glass
Glass Dataset
image
Image Dataset
leukemia
Leukemia Dataset
lymphoma
Lymphoma Dataset
NCI60
NCI60 Dataset
olive
Olive Dataset
parkinson
Parkinson Dataset
wine
Wine Dataset