build.nn: Builds a Neural Network object

In coldfir3/neatr: NeuroEvolution of Augmenting Topologies

Description

Builds a Neural Network object

Usage

build.nn(inputs, outputs, hiddens = 0, weights = NULL,
  weight_range = c(-1, 1), enable_rate = 0.7)

Arguments

inputs, outputs	Numeric indicating the number of inputs and outputs
hiddens	Numeric Vector indicating the number of hidden neurons per layer (optional)
weights	Numeric Vector of the nodal weight, if NULL (default) random weights will be sampled from an uniform distribution over weight_range. missing values will be randomized and excess values will be cropped
weight_range	Numeric Vector of length 2 with the lower and upper bounds for the random weight generation
enable_rate	Numeric between 0 and 1 indicating the probability that a link will be enabled

Value

neural network object

Examples

build.nn(2, 1, enable_rate = 1) %>% plot
build.nn(2, 1, 2, enable_rate = 1) %>% plot
build.nn(2, 1, c(2,3), enable_rate = 1, weight_range = 0) %>% plot
build.nn(2, 2, c(5,5,5), enable_rate = 1, weight_range = 0) %>% plot

coldfir3/neatr documentation built on May 23, 2019, 4:24 p.m.