plotNN: Plotting fitted neural networks

In gamlss.add: Extra Additive Terms for Generalized Additive Models for Location Scale and Shape

Description

A function to plot the results of a neural network fit based on the plotnet() function of the package NeuralNetTools

Usage

## S3 method for class 'nnet'
## S3 method for class 'nnet'
plot(x, nid = TRUE, all.out = TRUE, all.in = TRUE, bias = TRUE, 
wts.only = FALSE, rel.rsc = 5, circle.cex = 5, node.labs = TRUE, 
var.labs = TRUE, x.lab = NULL, y.lab = NULL, line.stag = NULL, 
struct = NULL, cex.val = 1, alpha.val = 1, circle.col = "lightblue",
pos.col = "black", neg.col = "grey", max.sp = FALSE, ...)

Arguments

x	A neural network fitted model
nid	logical value indicating if neural interpretation diagram is plotted, default is codeTRUE
all.out	character string indicating names of response variables for which connections are plotted, default all
all.in	character string indicating names of input variables for which connections are plotted, default all
bias	logical value indicating if bias nodes and connections are plotted, not applicable for networks from mlp function, default TRUE
wts.only	logical value indicating if connections weights are returned rather than a plot, default FALSE
rel.rsc	numeric value indicating maximum width of connection lines, default 5
circle.cex	numeric value indicating size of nodes, passed to cex argument, default 5
node.labs	logical value indicating if text labels are plotted, default TRUE
var.labs	logical value indicating if variable names are plotted next to nodes, default TRUE
x.lab	character string indicating names for input variables, default from model object
y.lab	character string indicating names for output variables, default from model object
line.stag	numeric value that specifies distance of connection weights from nodes
struct	numeric value of length three indicating network architecture (no nodes for input, hidden, output), required only if mod.in is a numeric vector
cex.val	numeric value indicating size of text labels, default 1
alpha.val	numeric value (0-1) indicating transparency of connections, default 1
circle.col	text value indicating colour of nodes default "lighrblue"
pos.col	text value indicating colour of the possitive connections, default "black"
neg.col	text value indicating colour of the negative connections, default "gray"
max.sp	logical value indication whether the space betwwen nodes in each laers is maximised
...	for further arguments

Details

The function plot.nnet() is (almost) identical to the function plot.nnet() created by Marcus W. Beck it was first published in the web but now is part of the NeuralNetTools package in R under the name plotnet(). Here we modify the function it so it works within the gamlss.add package. This involves of borrowing the functions rescale(), zero_range() and alpha() from package scales.

Value

The function is producing a plot

Author(s)

Marcus W. Beck <[e-mail]> modified by Mikis Stasinopoulos

References

Marcus W. Beck (2015). NeuralNetTools: Visualization and Analysis Tools for Neural Networks. R package version 1.4.1. https://cran.r-project.org/package=NeuralNetTools

Hadley Wickham (2014). scales: Scale functions for graphics. R package version 0.4.0. https://cran.r-project.org/package=scales

See Also

nn

Examples

r1 <- gamlss(R~nn(~Fl+A+H+loc, size=10, decay=0.2), data=rent, 
      family=GA, gd.tol=1000, n.cyc=5)
getSmo(r1)
plot(getSmo(r1), y.lab=expression(eta[1]))
plot(getSmo(r1), y.lab=expression(g[1](mu)))
## Not run: 
r2 <- gamlss(R~nn(~Fl+A+H+loc, size=10, decay=0.2), 
      sigma.fo=~nn(~Fl+A+H+loc, size=10, decay=0.2),data=rent, 
      family=GA, gd.tol=1000, n.cyc=5)
plot(getSmo(r2), y.lab=expression(g[1](mu)))
plot(getSmo(r2, what="sigma"), y.lab=expression(g[2](sigma)))

## End(Not run)

Example output

Loading required package: gamlss.dist
Loading required package: MASS
Loading required package: gamlss
Loading required package: splines
Loading required package: gamlss.data

Attaching package: 'gamlss.data'

The following object is masked from 'package:datasets':

    sleep

Loading required package: nlme
Loading required package: parallel
 **********   GAMLSS Version 5.1-3  ********** 
For more on GAMLSS look at http://www.gamlss.org/
Type gamlssNews() to see new features/changes/bug fixes.

Loading required package: mgcv
This is mgcv 1.8-28. For overview type 'help("mgcv-package")'.
Loading required package: nnet

Attaching package: 'nnet'

The following object is masked from 'package:mgcv':

    multinom

Loading required package: rpart
GAMLSS-RS iteration 1: Global Deviance = 27771.07 
GAMLSS-RS iteration 2: Global Deviance = 27756.47 
GAMLSS-RS iteration 3: Global Deviance = 27756.47 
a 5-10-1 network with 71 weights
inputs: Fl A H1 loc2 loc3 
output(s): y 
options were - linear output units  decay=0.2
GAMLSS-RS iteration 1: Global Deviance = 27575.54 
GAMLSS-RS iteration 2: Global Deviance = 27508.11 
GAMLSS-RS iteration 3: Global Deviance = 27492.5 
GAMLSS-RS iteration 4: Global Deviance = 27491.17 
GAMLSS-RS iteration 5: Global Deviance = 27491.17 

gamlss.add documentation built on Feb. 4, 2020, 9:08 a.m.