select.activation.function {AMORE}R Documentation

Provides R code of the selected activation function.

Description

Provides random values to the neuron weights and bias so as to start with. It is usually called by the random.init.NeuralNet function during the construction of the neural object by the newff function.

Usage

 select.activation.function(activation.function)

Arguments

activation.function

Activation function name. Currently the user may choose amongst purelin, tansig, sigmoid, hardlim and custom. If custom is chosen the the user must manually assign the neuron f0 and f1 functions.

Value

select.activation.function returns a list with two elements. The first, f0 is the R code selected to serve as the neuron activation function. The second, f1 is the R code of the activation function derivative.

Author(s)

Manuel Castejón Limas. manuel.castejon@gmail.com
Joaquin Ordieres Meré. j.ordieres@upm.es
Ana González Marcos. ana.gonzalez@unirioja.es
Alpha V. Pernía Espinoza. alpha.pernia@unirioja.es
Francisco Javier Martinez de Pisón. fjmartin@unirioja.es
Fernando Alba Elías. fernando.alba@unavarra.es

See Also

init.MLPneuron, newff


[Package AMORE version 0.2-15 Index]