shipunov (version 1.0)

plot.nnet: Plots 'nnet' object

Description

Plots 'nnet' object

Usage

# S3 method for nnet
plot(x, ..., nid=TRUE, all.out=TRUE, all.in=TRUE, wts.only=FALSE,
 rel.rsc=5, circle.cex=5, node.labs=TRUE, line.stag=NULL, cex.val=1, alpha.val=1,
 circle.col="lightgrey", pos.col="black", neg.col="grey")

Arguments

x

'nnet' model

...

arguments to 'plot()'

nid

TRUE

all.out

TRUE

all.in

TRUE

wts.only

FALSE

rel.rsc

5

circle.cex

5

node.labs

TRUE

line.stag

NULL

cex.val

1

alpha.val

1

circle.col

"lightgrey"

pos.col

"black"

neg.col

"grey"

Details

Plots 'nnet' object. Requires 'nnet' and 'scales' libraries. Based on the code from Marcus W Beck.

Examples

# NOT RUN {
library(nnet)
iris.train <- iris[seq(1, nrow(iris), 5), ]
iris.unknown <- iris[-seq(1, nrow(iris), 5), ]
iris.nnet <- nnet(Species ~ . , data=iris.train, size=4)
iris.predicted <- predict(iris.nnet, iris.unknown[, -5], type="class")
Misclass(iris.predicted, iris.unknown[, 5])
##
oldpar <- par(mar=c(0, 0, 0, 0))
plot(iris.nnet)
par(oldpar)
# }