Train models without tuning for performance

flash_models(d, outcome, models = c("rf", "knn"), metric, positive_class,
  n_folds = 5, model_class)

Arguments

d

A data frame

outcome

Name of the column to predict

models

Names of models to try, by default "rf" for random forest and "knn" for k-nearest neighbors. See supported_models for available models.

metric

What metric to use to assess model performance? Options for regression: "RMSE" (root-mean-squared error, default), "MAE" (mean-absolute error), or "Rsquared." For classification: "ROC" (area under the receiver operating characteristic curve), or "PR" (area under the precision-recall curve).

positive_class

For classification only, which outcome level is the "yes" case, i.e. should be associated with high probabilities? Defaults to "Y" or "yes" if present, otherwise is the first level of the outcome variable (first alphabetically if the training data outcome was not already a factor).

n_folds

How many folds to train the model on. Default = 5, minimum = 2. Whie flash_models doesn't use cross validation to tune hyperparameters, it trains n_folds models to evaluate performance out of fold.

model_class

"regression" or "classification". If not provided, this will be determined by the class of `outcome` with the determination displayed in a message.

Value

A model_list object. You can call plot, summary, evaluate, or predict on a model_list.

Details

This function has two major differences from tune_models: 1. It uses fixed default hyperparameter values to train models instead of using cross-validation to optimize hyperparameter values for predictive performance, and, as a result, 2. It is much faster.

If you want to train a model at a single set of non-default hyperparameter values use tune_models and pass a single-row data frame to the hyperparameters arguemet.

See also

For setting up model training: prep_data, supported_models, hyperparameters

For evaluating models: plot.model_list, evaluate.model_list

For making predictions: predict.model_list

For optimizing performance: tune_models

To prepare data and tune models in a single step: machine_learn

Examples

# NOT RUN {
# Prepare data
prepped_data <- prep_data(pima_diabetes, patient_id, outcome = diabetes)

# Get models quickly at default hyperparameter values
flash_models(prepped_data, diabetes)

# Speed comparison of no tuning with flash_models vs. tuning with tune_models:
# ~15 seconds:
system.time(
  tune_models(prepped_data, diabetes)
)
# ~3 seconds:
system.time(
  flash_models(prepped_data, diabetes)
)
# }