Get EvaluatedList Object Summary

This function produces a summary of an object of class EvaluatedList.

Usage

summary_table.EvaluatedList(object, key, best = c("opt", "1se"), ...)

object: object of class EvaluatedList, a list of Evaluated elements
key: (optional) character vector indicating the data to use for uniquely identifying the evaluated elements from the list. It can be a combination of "id", "config", "response", and "sampling". If missing, it is internally identified with identify_primary_key_from_EvaluatedList

A data.frame with the following elements

name: character string indicating a name for the considered settings. It is generated by generate_name
config: name of the selected tuned configuration
response: response type
sampling: sampling strategy
training_set_size: training-set size
k: number of iteration of the sampling procedure
tuning: character string, the tuning strategy
screening: character string, the screening strategy
screened: integer, the number of screened features
learning: character string, the learning method
nfeatures: integer, the number of features in the model
configuration: string, the configuration, i.e. the fixed hyperparameters of the model
imodel: index of the model
set: name of the set
measure: name of the considered performance metric
score: performance metrics
mean_score: mean performance metric
standard_error: standard error of the mean performance metric
upper_ci: upper bound of confidence interval
lower_ci: lower bound of confidence interval
best_model: logical indicating which is the best model
selection_rule: character string indicating the rule adopted for selection of tuned/trained model across k runs

Alessandro Barberis