find_best_model: Find the best version of each algorithm based on the...

In snfagora/autotextclassifier: R Package for Automatically Classifying Texts Based on Tidymodels

Find the best version of each algorithm based on the hyperparameters and 10-fold cross-validation.

Description

Find the best version of each algorithm based on the hyperparameters and 10-fold cross-validation.

Usage

find_best_model(
  lasso_wf,
  rand_wf,
  xg_wf,
  class_folds,
  lasso_grid,
  rand_grid,
  xg_grid,
  metric_choice = "accuracy"
)

Arguments

lasso_wf	A lasso workflow (including the search space for the model)
rand_wf	A random forest workflow (including the search space for the model)
xg_wf	An XGBoost workflow (including the search space for the model)
class_folds	10-fold cross-validation samples
lasso_grid	The search spaces for lasso
rand_grid	The search spaces for random forest
xg_grid	The search space for XGBoost
metric_choice	The selected metrics for the model evaluation among accuracy, balanced accuracy (bal_accuracy), F-score (f_means), and Area under the ROC curve (roc_auc). The default value is accuracy.

Value

A list output that contains the best model output for lasso, random forest, and XGBoost.

snfagora/autotextclassifier documentation built on June 16, 2022, 9:42 a.m.