Building and customizing ggplot2-based diagnostic plots with the mixpoissonreg package
ggplot2-based diagnostic plots
To obtain ggplot2-based diagnostic plots for
mixpoissonreg objects one should use the
autoplot method.
In the subsequent sections we will show how to use the
autoplot method as well as how to use its arguments to
customize the plots.
Choosing plots
The first, and maybe the easiest, customizable option is the choice of plots to be displayed. The current available plots are (together with their corresponding numbers):
- Residuals vs Obs. number
- Normal Q-Q
- Cook’s distance
- Generalized Cook’s distance
- Cook’s dist vs Generalized Cook’s dist
- Response vs Fitted means
These plots can be chosen by giving a list of the numbers of the
wanted plots in the which argument.
If the argument which is not provided, then, by default,
the plots 1, 2, 5 and 6 will be displayed:
library(mixpoissonreg)
fit <- mixpoissonreg(daysabs ~ gender + math + prog | gender + math + prog,
data = Attendance)
autoplot(fit)
If we want only a single plot, we simply indicate its number. For
instance, if we only want the plot of Cook’s distances, we simply set
which = 3:
If we want more than one, we provide a list with the desired plots.
Suppose we want the global influence-related plots, that is, plots 3, 4
and 5. Then, we set which = c(3,4,5):
Titles and subtitles
In this section we will describe how to customize titles and subtitles.
First of all, by default, the type of the fitted model (that is, if it is a Negative-Binomial or Poisson Inverse Gaussian regression) is included in each title. See, for instance, the plot below:
To remove the model type, simply set include.modeltype
to FALSE:
One should notice that in both plots, the type of the residual was
only included in the y-axis label. To also include the type of
the residual at the title, simply set include.residualtype
to TRUE:
The titles of the plots (displayed above the plot) can be altered by
setting the title parameter to a list containing the wanted
titles. One drawback is that we have to provide a list containing the
titles we want in the positions of the plots we want. For example, if we
only have one plot, the following example will change the caption:
Similarly, if we wanted to set the titles of plots 2 and 4, we could do:
my_titles <- rep("",6)
my_titles[2] <- "My title 2"
my_titles[4] <- "My title 4"
autoplot(fit, which = c(2,4), title = my_titles)
Notice that, since we did not change the
include.modeltype argument, the model type are added in the
new captions.
We can change the size of the titles with the argument
title.size. So, for instance,
We can also turn the title bold by setting the argument
title.bold to TRUE:
We can also change the subcaption. By default, the subcaption is a
simplified version of the call to the mixpoissonreg
function that was used to fit the model.
We must have some caution on describing the subcaption. If each plot
is given in one window (in the case both ncol and
nrow are NULL), then the subcaption is the
caption below the x-axis label. However, if multiple plots are
given at once, and there is space on the upper part of the plot, then
the subcaption is a general caption for all the plots.
We will illustrate the above description with examples to make it clearer.
The subcaption can be altered by setting the sub.caption
parameter to the desired caption.
Thus, we begin by providing the situation of one plot at a time.
Notice that the subcaption is below the x-axis label. The
same happens even if we there is more than one plot but both
ncol and nrow are NULL:
Now, notice the position of the subcaption when we gather multiple
plots using the one of the arguments ncol and/or
nrow:
In the previous case, that is, the case in which we have the
subcaption above all plots,notice that the common title is bold by
default. One can make it not bold by setting
gpar_sub.caption argument to list():
Similarly, we can change the size of the subcaption using the
argument by defining a fontsize element to the
gpar_sub.caption list. For instance,
autoplot(fit, sub.caption = "My subcaption", nrow = 2,
gpar_sub.caption = list(fontface = "bold", fontsize = 30))
Colors, sizes, point types and line types
In this section we show how to customize colors, sizes and types of lines and points.
First of all, let us change the shape of the points. To this end we
set the shape argument to the desired shape:
Now let us deal with the point colors. To change the colors of the
points, we use the colour argument:
To change the point sizes, we use the size argument:
We will now deal with point labels. First of all, we may change the
color of the point labels by using the argument
label.colour:
We can change the point labels’ sizes by using the
label.size argument:
Another example is the title color. We can change the title color
with the title.colour argument:
We can also change the title size, by using the
title.size argument:
Similarly, we can change the x and y label’s colors
by using the x.axis.col and y.axis.col
arguments:
As well as change their sizes by using the x.axis.size
and y.axis.size arguments:
We can change the sub.caption color by using the
sub.caption.col argument:
and change its size by using the sub.caption.size
argument:
We can also change the font face by using the
sub.caption.face argument:
We can change the position of the subcaption by changing the
sub.caption.hjust argument. The default is 0.5 which
indicates that the subcaption is centered. We can place the subcaption
to the left side of the plot by setting it to 0:
We can also place it on the right side of the plot by setting it to 1:
Let us now customize the lines in the Cook’s distance plots, namely plots 3 and 4. By default, the line type for Cook’s distance plots is “linerange”:
We change the line colors by using the colour
argument:
Let us change the line type for Cook’s distance plots to points and change the point types to crosses:
Finally, let us customize the quantile-quantile plots with and without simulated envelopes, namely, plot 2. The first customization is to remove the diagonal Q-Q line in the quantile-quantile plot without simulated envelopes:
We can change the qqline color by using the ad.colour
argument:
Finally, let us consider a fitting with simulated envelopes:
fit_env <- mixpoissonregML(daysabs ~ gender + math + prog | gender +
math + prog, envelope = 100, data = Attendance)
autoplot(fit_env, which = 2)
Let us first change the color of the median line of the simulated envelopes:
Let us also change the fill color of the simulated envelopes:
We can change the transparency of the envelope by changing the
env_alpha argument:
Points to be identified
By default the autoplot.mixpoissonreg function always
identifies the 3 “most extreme” points. We can change it so that it does
not identify any points by setting the argument label.n to
0:
We can also increase the number of identified points. For instance:
Finally, we can change the labels of the identified points with the
argument label.label. For instance, we may want to have the
value of the prog covariate instead:
Multiple plots
We can easily create frames of multiple plots by simply using one of
the arguments nrow or ncol. For instance,
As previously mentioned, we can change the common title in a multiple
plot by changing the sub.caption argument:
To customize the common title, one can provide a list containing the
“customizations” to the gpar_sub.caption argument. For
instance:
autoplot(fit, nrow = 2, sub.caption = "Mixed Poisson Regression",
gpar_sub.caption = list(fontsize = 30, col = "red",
fontface = "italic"))