This function builds a classification model using Multilayer Perceptron.
MLP(
train,
labels,
hidden = if (is.vector(train)) 2:(1 + nlevels(labels)) else 2:(ncol(train) +
nlevels(labels)),
decay = 10^(-3:-1),
nfolds = 10,
tune = FALSE,
methodparameters = NULL,
graph = FALSE,
seed = NULL,
...
)The classification model.
The training set (description), as a data.frame.
Class labels of the training set (vector or factor).
The size of the hidden layer (if a vector, cross-over validation is used to chose the best size).
The decay (between 0 and 1) of the backpropagation algorithm (if a vector, cross-over validation is used to chose the best size).
The number of folds of the cross-validation a method runs to choose its hyperparameters. Only used when there is something to choose, i.e. when one of them is given as a vector. Lower it to fit faster, at the cost of a noisier choice.
If true, the function returns parameters instead of a classification model.
Object containing the parameters. If given, it replaces size and decay.
Present for interface consistency with performance (which always
passes it when fitting a model). Currently unused: MLP does not produce a plot.
A specified seed for random number generation, so that two runs on the same data give the same model. Every learning method accepts it, so that it can be set the same way whatever the method; the deterministic ones simply have nothing to draw and give the same model with or without it.
Other parameters.
# \donttest{
require (datasets)
data (iris)
MLP (iris [, -5], iris [, 5], hidden = 4, decay = .1)
# }
Run the code above in your browser using DataLab