fdm2id (version 0.9.5)

MLPREG: Multi-Layer Perceptron Regression

Description

This function builds a regression model using MLP.

Usage

MLPREG(
  x,
  y,
  size = 2:(ifelse(is.vector(x), 2, ncol(x))),
  decay = 10^(-3:-1),
  params = NULL,
  tune = FALSE,
  ...
)

Arguments

x

Predictor matrix.

y

Response vector.

size

The size of the hidden layer (if a vector, cross-over validation is used to chose the best size).

decay

The decay (between 0 and 1) of the backpropagation algorithm (if a vector, cross-over validation is used to chose the best size).

params

Object containing the parameters. If given, it replaces size and decay.

tune

If true, the function returns paramters instead of a classification model.

...

Other parameters.

Value

The classification model, as an object of class model-class.

See Also

nnet

Examples

# NOT RUN {
require (datasets)
data (trees)
MLPREG (trees [, -3], trees [, 3])
# }