| 42333 |
ripley |
1 |
% File src/library/graphics/man/cdplot.Rd
|
|
|
2 |
% Part of the R package, http://www.R-project.org
|
| 59039 |
ripley |
3 |
% Copyright 1995-2007 R Core Team
|
| 42333 |
ripley |
4 |
% Distributed under GPL 2 or later
|
|
|
5 |
|
| 35517 |
hornik |
6 |
\name{cdplot}
|
| 56186 |
murdoch |
7 |
\alias{cdplot}
|
| 35517 |
hornik |
8 |
\alias{cdplot.default}
|
|
|
9 |
\alias{cdplot.formula}
|
|
|
10 |
\title{Conditional Density Plots}
|
|
|
11 |
\description{
|
|
|
12 |
Computes and plots conditional densities describing how the
|
|
|
13 |
conditional distribution of a categorical variable \code{y} changes over a
|
|
|
14 |
numerical variable \code{x}.
|
|
|
15 |
}
|
|
|
16 |
\usage{
|
|
|
17 |
cdplot(x, \dots)
|
| 36139 |
ripley |
18 |
|
| 35517 |
hornik |
19 |
\method{cdplot}{default}(x, y,
|
| 41355 |
hornik |
20 |
plot = TRUE, tol.ylab = 0.05, ylevels = NULL,
|
| 35517 |
hornik |
21 |
bw = "nrd0", n = 512, from = NULL, to = NULL,
|
|
|
22 |
col = NULL, border = 1, main = "", xlab = NULL, ylab = NULL,
|
| 35599 |
hornik |
23 |
yaxlabels = NULL, xlim = NULL, ylim = c(0, 1), \dots)
|
| 36139 |
ripley |
24 |
|
| 35517 |
hornik |
25 |
\method{cdplot}{formula}(formula, data = list(),
|
| 41355 |
hornik |
26 |
plot = TRUE, tol.ylab = 0.05, ylevels = NULL,
|
| 35517 |
hornik |
27 |
bw = "nrd0", n = 512, from = NULL, to = NULL,
|
|
|
28 |
col = NULL, border = 1, main = "", xlab = NULL, ylab = NULL,
|
| 35600 |
hornik |
29 |
yaxlabels = NULL, xlim = NULL, ylim = c(0, 1), \dots,
|
|
|
30 |
subset = NULL)
|
| 35517 |
hornik |
31 |
}
|
|
|
32 |
\arguments{
|
| 52994 |
hornik |
33 |
\item{x}{an object, the default method expects a single numerical
|
|
|
34 |
variable (or an object coercible to this).}
|
| 35517 |
hornik |
35 |
\item{y}{a \code{"factor"} interpreted to be the dependent variable}
|
|
|
36 |
\item{formula}{a \code{"formula"} of type \code{y ~ x} with a single dependent
|
| 61433 |
ripley |
37 |
\code{"factor"} and a single numerical explanatory variable.}
|
| 35517 |
hornik |
38 |
\item{data}{an optional data frame.}
|
|
|
39 |
\item{plot}{logical. Should the computed conditional densities be plotted?}
|
|
|
40 |
\item{tol.ylab}{convenience tolerance parameter for y-axis annotation.
|
|
|
41 |
If the distance between two labels drops under this threshold, they are
|
|
|
42 |
plotted equidistantly.}
|
| 41355 |
hornik |
43 |
\item{ylevels}{a character or numeric vector specifying in which order
|
|
|
44 |
the levels of the dependent variable should be plotted.}
|
| 35517 |
hornik |
45 |
\item{bw, n, from, to, \dots}{arguments passed to \code{\link{density}}}
|
|
|
46 |
\item{col}{a vector of fill colors of the same length as \code{levels(y)}.
|
| 61433 |
ripley |
47 |
The default is to call \code{\link{gray.colors}}.}
|
|
|
48 |
\item{border}{border color of shaded polygons.}
|
| 35517 |
hornik |
49 |
\item{main, xlab, ylab}{character strings for annotation}
|
| 35533 |
hornik |
50 |
\item{yaxlabels}{character vector for annotation of y axis, defaults to
|
|
|
51 |
\code{levels(y)}.}
|
| 35599 |
hornik |
52 |
\item{xlim, ylim}{the range of x and y values with sensible defaults.}
|
| 35600 |
hornik |
53 |
\item{subset}{an optional vector specifying a subset of observations
|
|
|
54 |
to be used for plotting.}
|
| 35517 |
hornik |
55 |
}
|
|
|
56 |
\details{
|
|
|
57 |
\code{cdplot} computes the conditional densities of \code{x} given
|
|
|
58 |
the levels of \code{y} weighted by the marginal distribution of \code{y}.
|
|
|
59 |
The densities are derived cumulatively over the levels of \code{y}.
|
|
|
60 |
|
|
|
61 |
This visualization technique is similar to spinograms (see \code{\link{spineplot}})
|
|
|
62 |
and plots \eqn{P(y | x)} against \eqn{x}. The conditional probabilities
|
| 40281 |
ripley |
63 |
are not derived by discretization (as in the spinogram), but using a smoothing
|
| 35517 |
hornik |
64 |
approach via \code{\link{density}}.
|
| 61433 |
ripley |
65 |
|
|
|
66 |
Note, that the estimates of the conditional densities are more reliable for
|
| 35517 |
hornik |
67 |
high-density regions of \eqn{x}. Conversely, the are less reliable in regions
|
|
|
68 |
with only few \eqn{x} observations.
|
|
|
69 |
}
|
|
|
70 |
\value{
|
|
|
71 |
The conditional density functions (cumulative over the levels of \code{y})
|
|
|
72 |
are returned invisibly.
|
|
|
73 |
}
|
|
|
74 |
\seealso{
|
|
|
75 |
\code{\link{spineplot}}, \code{\link{density}}
|
|
|
76 |
}
|
|
|
77 |
\references{
|
|
|
78 |
Hofmann, H., Theus, M. (2005), \emph{Interactive graphics for visualizing
|
|
|
79 |
conditional distributions}, Unpublished Manuscript.
|
|
|
80 |
}
|
|
|
81 |
\author{
|
|
|
82 |
Achim Zeileis \email{Achim.Zeileis@R-project.org}
|
|
|
83 |
}
|
|
|
84 |
\examples{
|
|
|
85 |
## NASA space shuttle o-ring failures
|
| 44236 |
ripley |
86 |
fail <- factor(c(2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 2, 1, 2, 1, 1, 1,
|
|
|
87 |
1, 2, 1, 1, 1, 1, 1),
|
| 35517 |
hornik |
88 |
levels = 1:2, labels = c("no", "yes"))
|
| 44236 |
ripley |
89 |
temperature <- c(53, 57, 58, 63, 66, 67, 67, 67, 68, 69, 70, 70,
|
|
|
90 |
70, 70, 72, 73, 75, 75, 76, 76, 78, 79, 81)
|
| 35517 |
hornik |
91 |
|
|
|
92 |
## CD plot
|
|
|
93 |
cdplot(fail ~ temperature)
|
|
|
94 |
cdplot(fail ~ temperature, bw = 2)
|
|
|
95 |
cdplot(fail ~ temperature, bw = "SJ")
|
|
|
96 |
|
|
|
97 |
## compare with spinogram
|
|
|
98 |
(spineplot(fail ~ temperature, breaks = 3))
|
|
|
99 |
|
| 41355 |
hornik |
100 |
## highlighting for failures
|
|
|
101 |
cdplot(fail ~ temperature, ylevels = 2:1)
|
|
|
102 |
|
| 35517 |
hornik |
103 |
## scatter plot with conditional density
|
|
|
104 |
cdens <- cdplot(fail ~ temperature, plot = FALSE)
|
|
|
105 |
plot(I(as.numeric(fail) - 1) ~ jitter(temperature, factor = 2),
|
|
|
106 |
xlab = "Temperature", ylab = "Conditional failure probability")
|
|
|
107 |
lines(53:81, 1 - cdens[[1]](53:81), col = 2)
|
|
|
108 |
}
|
|
|
109 |
\keyword{hplot}
|
The R Project SVN
R