Tuner Class — Tuner-class • renoir

An S4 class providing the methods to perform the tuning of hyperparameters and the training of the final model.

The object consists of 6 slots

Constructor for the S4 Tuner object.

Usage

Tuner(
  id = "grid.search",
  tuner,
  sampler = Sampler(method = "cv", k = 10L),
  looper = Looper(),
  screener = ScreenerList(Screener(id = "ebayes")),
  logger = Logger(),
  parameters
)

Arguments

id

object identifier, the name of the tuning method to use. If id is one of the supported by renoir, the constructor will automatically select a built-in tuner function.

tuner

(optional) function for tuning the hyperparameters of the model If tuning method is one of the supported by renoir, the constructor will automatically select a built-in tuner function. If tuner is provided it must conform to the renoir common interface, and must have the following formal arguments:

sampler: a Sampler object, the sampling procedure to use for tuning
looper: a Looper object
screener: a ScreenerList object, can contain just one Screener
trainer: a Trainer object
tester: a Tester object
selector: a Selector object
logger: a Logger object
hyperparameters: the input matrix, where rows are observations and columns are variables
x: the input matrix, where rows are observations and columns are variables
y: the response variable. Its number of rows must match the number of rows of x
weights: priors of the observations
offset: a priori known component to be included in the linear predictor during fitting
resp.type: the response type
...: additional arguments

It must return a Tuned or TunedList object

sampler

a Sampler object, the sampling procedure to use for tuning

looper

a Looper object

screener

a Screener object

logger

a Logger object

parameters

list containing the parameters to fix for the chosen tuning method. Please, note that the parameters shouldn't include the default renoir formals as they won't be considered. (It is helpful to change default settings)

Value

An object of class Tuner.

Slots

id: the name of the tuning method to use
tuner: function to use for tuning the hyperparameters of the model
parameters: list containing the parameters to fix for the chosen tuning method
logger: a Logger
sampler: a Sampler object, the sampling procedure to use for tuning
looper: a Looper object
screener: a ScreenerList object

Author

Alessandro Barberis