iterator: Run iterator

View source: R/iterator_class.R

runR Documentation

Run iterator

Description

Runs an iterator, applying the chosen model multiple times.

Evaluates an iterator by e.g. averaging over all iterations. May be deprecated in a future release as evaluate is applied by run anyway.

A class for iterative approaches that involve the training/prediction of a model multiple times. Not intended to be called directly, this class should be inherited to provide functionality for method-specific classes.

Usage

run(I, D, MET)

evaluate(I, MET)

iterator(...)

## S4 method for signature 'iterator,DatasetExperiment,metric'
run(I, D, MET = NULL)

## S4 method for signature 'iterator,metric'
evaluate(I, MET)

## S4 method for signature 'iterator'
models(ML)

## S4 replacement method for signature 'iterator,model_OR_iterator'
models(ML) <- value

## S4 replacement method for signature 'iterator,character'
result_name(I) <- value

## S4 method for signature 'iterator'
result(M)

## S4 method for signature 'iterator'
result_name(M)

## S4 method for signature 'iterator,model_OR_iterator'
e1 * e2

## S4 method for signature 'iterator,ANY,ANY,ANY'
x[i]

## S4 replacement method for signature 'iterator,ANY,ANY,ANY'
x[i] <- value

Arguments

I

an iterator object

D

a dataset object

MET

a metric object

...

named slots and their values.

ML

a model sequence object

value

value

M

a model object

e1

an iterator object

e2

an iterator or a model object

x

a sequence object

i

index into sequence

Details

Running an iterator will apply the iterator a number of times to a dataset_ For example, in cross-validation the same model is applied multiple times to the same data, splitting it into training and test sets. The input metric object can be calculated and collected for each iteration as an output_

Value

Modified iterator object

Modified iterator object

the modified model object

model at the given index in the sequence

iterator with the model at index i replaced

Examples

D = iris_DatasetExperiment() # get some data
MET = metric()  # use a metric
I = example_iterator() # initialise iterator
models(I) = example_model() # set the model
I = run(I,D,MET) # run
D = iris_DatasetExperiment() # get some data
MET = metric()  # use a metric
I = example_iterator() # initialise iterator
models(I) = example_model() # set the model
I = run(I,D,MET) # run
I = evaluate(I,MET) # evaluate
I = iterator()
I = iterator() * model()
D = DatasetExperiment()
MET = metric()
I = iterator() * model()
I = run(I,D,MET)

I = iterator()
result_name(I) = 'example'
MS = model() + model()
I = iterator() * MS
I[2] # returns the second model() object

MS = model() + model()
I = iterator() * MS
I[2] = model() # sets the second model to model()


computational-metabolomics/struct documentation built on March 27, 2024, 4:26 p.m.