Function reference • tune

Functions for tuning

tune_bayes(): Bayesian optimization of model parameters.

tune_grid(): Model tuning via grid search

fit_resamples(): Fit multiple models via resampling

fit_best(): Fit a model to the numerically optimal configuration

last_fit(): Fit the final best model to the training set and evaluate the test set

prob_improve() exp_improve() conf_bound(): Acquisition function for scoring parameter combinations

control_bayes(): Control aspects of the Bayesian search process

control_last_fit(): Control aspects of the last fit process

autoplot(<tune_results>): Plot tuning search results

augment(<tune_results>) augment(<resample_results>) augment(<last_fit>): Augment data with holdout predictions

coord_obs_pred(): Use same scale for plots of observed vs predicted values

expo_decay(): Exponential decay function

collect_predictions() collect_metrics() collect_notes() collect_extracts(): Obtain and format results produced by tuning functions

compute_metrics(): Calculate and format metrics from tuning functions

filter_parameters(): Remove some tuning parameter results

show_best() select_best() select_by_pct_loss() select_by_one_std_err(): Investigate best tuning parameters

show_notes(): Display distinct errors from tune objects

extract_workflow(<last_fit>) extract_workflow(<tune_results>) extract_spec_parsnip(<tune_results>) extract_recipe(<tune_results>) extract_fit_parsnip(<tune_results>) extract_fit_engine(<tune_results>) extract_mold(<tune_results>) extract_preprocessor(<tune_results>): Extract elements of tune objects

extract_model(): Convenience functions to extract model

finalize_model() finalize_recipe() finalize_workflow(): Splice final parameters into objects

conf_mat_resampled(): Compute average confusion matrix across resamples

example_ames_knn ames_wflow ames_grid_search ames_iter_search: Example Analysis of Ames Housing Data

int_pctl(<tune_results>): Bootstrap confidence intervals for performance metrics

Developer functions

merge(<recipe>) merge(<model_spec>): Merge parameter grid values into objects

parameters(<workflow>) parameters(<model_spec>) parameters(<recipe>): Determination of parameter sets for other objects

message_wrap(): Write a message that respects the line width

.use_case_weights_with_yardstick(): Determine if case weights should be passed on to yardstick

.stash_last_result(): Save most recent results to search path

control_grid() control_resamples() new_backend_options(): Control aspects of the grid search process