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
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/ppd-intervals.R
\name{PPD-intervals}
\alias{PPD-intervals}
\alias{ppd_intervals}
\alias{ppd_intervals_grouped}
\alias{ppd_ribbon}
\alias{ppd_ribbon_grouped}
\alias{ppd_intervals_data}
\alias{ppd_ribbon_data}
\title{PPD intervals}
\usage{
ppd_intervals(
ypred,
x = NULL,
...,
prob = 0.5,
prob_outer = 0.9,
alpha = 0.33,
size = 1,
fatten = 2.5,
linewidth = 1
)
ppd_intervals_grouped(
ypred,
x = NULL,
group,
...,
facet_args = list(),
prob = 0.5,
prob_outer = 0.9,
alpha = 0.33,
size = 1,
fatten = 2.5,
linewidth = 1
)
ppd_ribbon(
ypred,
x = NULL,
...,
prob = 0.5,
prob_outer = 0.9,
alpha = 0.33,
size = 0.25
)
ppd_ribbon_grouped(
ypred,
x = NULL,
group,
...,
facet_args = list(),
prob = 0.5,
prob_outer = 0.9,
alpha = 0.33,
size = 0.25
)
ppd_intervals_data(
ypred,
x = NULL,
group = NULL,
...,
prob = 0.5,
prob_outer = 0.9
)
ppd_ribbon_data(
ypred,
x = NULL,
group = NULL,
...,
prob = 0.5,
prob_outer = 0.9
)
}
\arguments{
\item{ypred}{An \code{S} by \code{N} matrix of draws from the posterior (or prior)
predictive distribution. The number of rows, \code{S}, is the size of the
posterior (or prior) sample used to generate \code{ypred}. The number of
columns, \code{N}, is the number of predicted observations.}
\item{x}{A numeric vector to use as the x-axis
variable. For example, \code{x} could be a predictor variable from a
regression model, a time variable for time-series models, etc. If \code{x}
is missing or \code{NULL} then the observation index is used for the x-axis.}
\item{...}{Currently unused.}
\item{prob, prob_outer}{Values between \code{0} and \code{1} indicating the desired
probability mass to include in the inner and outer intervals. The defaults
are \code{prob=0.5} and \code{prob_outer=0.9}.}
\item{alpha, size, fatten, linewidth}{Arguments passed to geoms. For ribbon
plots \code{alpha} is passed to \code{\link[ggplot2:geom_ribbon]{ggplot2::geom_ribbon()}} to control the opacity
of the outer ribbon and \code{size} is passed to \code{\link[ggplot2:geom_path]{ggplot2::geom_line()}} to
control the size of the line representing the median prediction (\code{size=0}
will remove the line). For interval plots \code{alpha}, \code{size}, \code{fatten}, and
\code{linewidth} are passed to \code{\link[ggplot2:geom_linerange]{ggplot2::geom_pointrange()}} (\code{fatten=0} will
remove the point estimates).}
\item{group}{A grouping variable of the same length as \code{y}.
Will be coerced to \link[base:factor]{factor} if not already a factor.
Each value in \code{group} is interpreted as the group level pertaining
to the corresponding observation.}
\item{facet_args}{A named list of arguments (other than \code{facets}) passed
to \code{\link[ggplot2:facet_wrap]{ggplot2::facet_wrap()}} or \code{\link[ggplot2:facet_grid]{ggplot2::facet_grid()}}
to control faceting. Note: if \code{scales} is not included in \code{facet_args}
then \strong{bayesplot} may use \code{scales="free"} as the default (depending
on the plot) instead of the \strong{ggplot2} default of \code{scales="fixed"}.}
}
\value{
The plotting functions return a ggplot object that can be further
customized using the \strong{ggplot2} package. The functions with suffix
\verb{_data()} return the data that would have been drawn by the plotting
function.
}
\description{
Medians and central interval estimates of posterior or prior predictive
distributions. Each of these functions makes the same plot as the
corresponding \code{\link[=PPC-intervals]{ppc_}} function but without plotting any
observed data \code{y}. The \strong{Plot Descriptions} section at \link{PPC-intervals} has
details on the individual plots.
}
\examples{
color_scheme_set("brightblue")
ypred <- example_yrep_draws()
x <- example_x_data()
group <- example_group_data()
ppd_intervals(ypred[, 1:50])
ppd_intervals(ypred[, 1:50], fatten = 0)
ppd_intervals(ypred[, 1:50], fatten = 0, linewidth = 2)
ppd_intervals(ypred[, 1:50], prob_outer = 0.75, fatten = 0, linewidth = 2)
# put a predictor variable on the x-axis
ppd_intervals(ypred[, 1:100], x = x[1:100], fatten = 1) +
ggplot2::labs(y = "Prediction", x = "Some variable of interest")
# with a grouping variable too
ppd_intervals_grouped(
ypred = ypred[, 1:100],
x = x[1:100],
group = group[1:100],
size = 2,
fatten = 0,
facet_args = list(nrow = 2)
)
# even reducing size, ppd_intervals is too cluttered when there are many
# observations included (ppd_ribbon is better)
ppd_intervals(ypred, size = 0.5, fatten = 0.1, linewidth = 0.5)
ppd_ribbon(ypred)
ppd_ribbon(ypred, size = 0) # remove line showing median prediction
}
\references{
Gabry, J. , Simpson, D. , Vehtari, A. , Betancourt, M. and
Gelman, A. (2019), Visualization in Bayesian workflow.
\emph{J. R. Stat. Soc. A}, 182: 389-402. doi:10.1111/rssa.12378.
(\href{https://rss.onlinelibrary.wiley.com/doi/full/10.1111/rssa.12378}{journal version},
\href{https://arxiv.org/abs/1709.01449}{arXiv preprint},
\href{https://github.com/jgabry/bayes-vis-paper}{code on GitHub})
}
\seealso{
Other PPDs:
\code{\link{PPD-distributions}},
\code{\link{PPD-overview}},
\code{\link{PPD-test-statistics}}
}
\concept{PPDs}
|
