Tuned Class — Tuned-class • renoir

An S4 class providing a container for the results of a tuning procedure.

Constructor for the S4 Tuned object.

Usage

Tuned(
  model = Trained(),
  config = character(),
  mscore = numeric(),
  sem = numeric(),
  tuning = character(),
  sampling = character(),
  measure = character(),
  screened = Screened(),
  xvars = integer(),
  xobs = integer(),
  index = integer(),
  trained = list(),
  validated = list(),
  stability = list()
)

Arguments

model: the tuned model
config: (optional) name of the configuration selected for the final model (e.g. 'opt' for configuration giving optimal accuracy during the repeated sampling procedure and '1se' for configuration giving least complex model having performance metric within 1SE from the optimal)
mscore: the average performance metric across the sampling procedure
sem: the standard error of the mscore
tuning: the tuning method
sampling: the sampling method used to tune the model
measure: the accuracy metric
screened: a Screened
xvars: the number of features in the input matrix
xobs: the number of observations in the input matrix
index: the index of the selected configuration
trained: a list containing the models trained during the resampling
validated: a list containing the validation of the trained models
stability: a list containing the features stability

Slots

model: a Trained object
config: a character string
mscore: a length-one numeric vector
sem: a length-one numeric vector
tuning: a character string
sampling: a character string
measure: a character string
screened: a Screened object
xvars: a length-one integer vector
xobs: a length-one integer vector
index: a length-one integer vector
trained: a list
validated: a list
stability: a list

Author

Alessandro Barberis