https://github.com/cran/neuralnet
Tip revision: c4e62da3b7aa50d399a9e1d112462b6f057873fd authored by Frauke Guenther on 10 August 2008, 00:00:00 UTC
version 1.2
version 1.2
Tip revision: c4e62da
compute.Rd
\name{compute}
\alias{compute}
\title{ Computation of a given neural network for a new covariate vector}
\description{
\code{compute}, a method for objects of class \code{nn}, typically
produced by \code{neuralnet}.
Computes the outputs of all neurons for a specific arbitrary covariate vector given a trained neural network. Please make sure that the order of the covariates is the same in the new matrix or dataframe as in the original neural network.
}
\usage{
compute(x, covariate, rep = 1)
}
\arguments{
\item{x}{ an object of class \code{nn}. }
\item{covariate}{ a data.frame or matrix containing the variables to calculate the output of the neural network. }
\item{rep}{ an integer indicating the neural network's repetition which should be used. }
}
\value{
\code{compute} returns a list containing the following components:
\item{neurons}{a list of the neuron's output for each layer of the neural network.}
\item{net.result}{a matrix containing the overall result of the neural network.}
}
\author{ Stefan Fritsch \email{fritsch@bips.uni-bremen.de} }
\examples{
Var1 <- runif(50, 0, 100)
sqrt.data <- data.frame(Var1, Sqrt=sqrt(Var1))
print(net.sqrt <- neuralnet( Sqrt~Var1, sqrt.data, hidden=10, threshold=0.01))
compute(net.sqrt, (1:10)^2)$net.result
}
\keyword{ neural }