1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
|
\name{partialPlot}
\alias{partialPlot}
\alias{partialPlot.default}
\alias{partialPlot.randomForest}
\title{Partial dependence plot}
\description{
Partial dependence plot gives a graphical depiction of the marginal
effect of a variable on the class probability (classification) or
response (regression).
}
\usage{
\method{partialPlot}{randomForest}(x, pred.data, x.var, which.class,
w, plot = TRUE, add = FALSE,
n.pt = min(length(unique(pred.data[, xname])), 51),
rug = TRUE, xlab=deparse(substitute(x.var)), ylab="",
main=paste("Partial Dependence on", deparse(substitute(x.var))),
...)
}
\arguments{
\item{x}{an object of class \code{randomForest}, which contains a
\code{forest} component.}
\item{pred.data}{a data frame used for contructing the plot, usually
the training data used to contruct the random forest.}
\item{x.var}{name of the variable for which partial
dependence is to be examined.}
\item{which.class}{For classification data, the class to focus on
(default the first class).}
\item{w}{weights to be used in averaging; if not supplied, mean is not
weighted}
\item{plot}{whether the plot should be shown on the graphic device.}
\item{add}{whether to add to existing plot (\code{TRUE}).}
\item{n.pt}{if \code{x.var} is continuous, the number of points on the
grid for evaluating partial dependence.}
\item{rug}{whether to draw hash marks at the bottom of the plot
indicating the deciles of \code{x.var}.}
\item{xlab}{label for the x-axis.}
\item{ylab}{label for the y-axis.}
\item{main}{main title for the plot.}
\item{...}{other graphical parameters to be passed on to \code{plot}
or \code{lines}.}
}
\value{
A list with two components: \code{x} and \code{y}, which are the values
used in the plot.
}
\details{
The function being plotted is defined as:
\deqn{
\tilde{f}(x) = \frac{1}{n} \sum_{i=1}^n f(x, x_{iC}),
}
where \eqn{x} is the variable for which partial dependence is sought,
and \eqn{x_{iC}} is the other variables in the data. The summand is
the predicted regression function for regression, and logits
(i.e., log of fraction of votes) for \code{which.class} for
classification:
\deqn{ f(x) = \log p_k(x) - \frac{1}{K} \sum_{j=1}^K \log p_j(x),}
where \eqn{K} is the number of classes, \eqn{k} is \code{which.class},
and \eqn{p_j} is the proportion of votes for class \eqn{j}.
}
\note{
The \code{randomForest} object must contain the \code{forest}
component; i.e., created with \code{randomForest(...,
keep.forest=TRUE)}.
This function runs quite slow for large data sets.
}
\references{
Friedman, J. (2001). Greedy function approximation: the gradient
boosting machine, \emph{Ann. of Stat.}}
\seealso{\code{\link{randomForest}}}
\author{Andy Liaw \email{andy_liaw@merck.com}}
\examples{
data(iris)
set.seed(543)
iris.rf <- randomForest(Species~., iris)
partialPlot(iris.rf, iris, Petal.Width, "versicolor")
## Looping over variables ranked by importance:
data(airquality)
airquality <- na.omit(airquality)
set.seed(131)
ozone.rf <- randomForest(Ozone ~ ., airquality, importance=TRUE)
imp <- importance(ozone.rf)
impvar <- rownames(imp)[order(imp[, 1], decreasing=TRUE)]
op <- par(mfrow=c(2, 3))
for (i in seq_along(impvar)) {
partialPlot(ozone.rf, airquality, impvar[i], xlab=impvar[i],
main=paste("Partial Dependence on", impvar[i]),
ylim=c(30, 70))
}
par(op)
}
\keyword{classif}
\keyword{regression}
\keyword{tree}
|
