build_keras_net: Build a Keras Multilayer Perceptron

In mlr3learners/mlr3learners.proba: Learners for mlr3proba

Description

Utility function to build a Keras MLP.

Usage

build_keras_net(
  n_in,
  n_out,
  nodes = c(32, 32),
  layer_pars = list(),
  activation = "linear",
  act_pars = list(),
  dropout = 0.1,
  batch_norm = TRUE,
  batch_pars = list()
)

Arguments

n_in	(integer(1)) Number of input features.
n_out	(integer(1)) Number of targets.
nodes	(numeric()) Hidden nodes in network, each element in vector represents number of hidden nodes in respective layer.
layer_pars	(list()) Arguments passed to keras::layer_dense.
activation	(character(1)) Activation function passed to keras::layer_activation. Default is linear.
act_pars	(list()) Parameters for activation function, see keras::layer_activation.
dropout	(numeric(1)) Optional dropout layer, if NULL then no dropout layer added otherwise either same dropout will be added to all layers.
batch_norm	(logical(1)) If TRUE (default) then batch normalisation is applied to all layers.
batch_pars	(list()) Parameters for batch normalisation, see keras::layer_batch_normalization.

Details

This function is a helper for R users with less Python experience. Currently it is limited to simple MLPs and with identical layers. More advanced networks will require manual creation with keras.

Examples

build_keras_net(10, 1)

build_keras_net(n_in = 10, n_out = 1, nodes = c(4, 4, 4), activation = "elu")

mlr3learners/mlr3learners.proba documentation built on Aug. 9, 2020, 12:50 a.m.