- model
a binary regression model fitted with glm
.
- main
user-specified main title.
- xlab
x-axis label. Defaults to the name of the (first) numeric predictor.
- ylab
y-axis label. Defaults to the name of the response -
within either 'P(...)' or 'logit(...)', depending on the response
type.
- xlim
Range of the x-axis. Defaults to the range of the numeric
predictor.
- ylim
Range of the y-axis. Defaults to the unit interval on
probability scale or the fitted values range on the link scale,
depending on type
.
- pred_var
character string of length 1 giving the name of
the numeric predictor. Defaults to the first one found in the data
set.
- pred_range
"data"
, "xlim"
, or a numeric vector.
If "data"
, the numeric predictor corresponds to the observed values. If
"xlim"
, 100 values are taken from the "xlim"
range. A numeric vector will be interpreted as the values to be predicted.
- group_vars
optional character string of conditioning
variables. Defaults to all factors found in the data set, response
excluded. If FALSE
, no variables are used for conditioning.
- base_level
vector of length one. If the response is a vector,
this specifies the base ('no effect') value of the
response variable
(e.g., "Placebo", 0, FALSE, etc.) and defaults
to the first level for
factor responses, or 0 for numeric/binary variables. This controls
which observations will be plotted on the top or the bottom of the
display. If the response is a matrix with success and failure
column, this specifies the one to be interpreted as failure
(default: 2), either as an integer, or as a
string ("success"
or "failure"
). The proportions of
successes will be plotted as observed values.
- subset
an optional vector specifying a subset of the data
rows. The value is evaluated in the data environment, so expressions
can be used to select the data (see examples).
- type
either "response" or "link" to select the scale of the
fitted values. The y-axis will be adapted accordingly.
- conf_level
confidence level used for calculating
confidence bands.
- delta
logical; indicates whether the delta method should be
employed for calculating the limits of the confidence band or not
(see details).
- pch
character or numeric vector of symbols used for plotting
the (possibly conditioned) observed values, recycled as needed.
- cex
size of the plot symbols (in lines).
- jitter_factor
argument passed to jitter
used for the points representing the observed values.
- lwd
Line width for the fitted values.
- lty
Line type for the fitted values.
- point_size
size of points for the fitted values in char units (default: 0, so
no points are plotted).
- col_lines, col_bands
character vector specifying the colors of the fitted
lines and confidence bands,
by default chosen with rainbow_hcl
. The
confidence bands are using alpha blending with alpha = 0.2.
- legend
logical; if TRUE
(default), a legend is drawn.
- legend_pos
numeric vector of length 2, specifying x and y
coordinates of the legend, or a character string (e.g., "topleft"
,
"center"
etc.). Defaults to "topleft"
if the fitted curve's slope is
positive, and "topright"
else.
- legend_inset
numeric vector or length 2 specifying the inset
from the legend's x and y coordinates in npc units.
- legend_vgap
vertical space between the legend's line entries.
- labels
logical; if TRUE
, labels corresponding to the
factor levels are plotted next to the fitted lines.
- labels_pos
either "right"
or "left"
, determining on which side
of the fitted lines (start or end) the labels should be placed.
- labels_just
character vector of length 2, specifying the
relative justification of the labels to their coordinates. See the
documentation of the just
parameter of
grid.text
for more details.
- labels_offset
numeric vector of length 2, specifying the offset
of the labels' coordinates in npc units.
- gp_main
object of class "gpar"
used for the main title.
- gp_legend_frame
object of class "gpar"
used for the
legend frame.
- gp_legend_title
object of class "gpar"
used for the
legend title.
- newpage
logical; if TRUE
, the plot is drawn on a new page.
- pop
logical; if TRUE
, all newly generated viewports are
popped after plotting.
- return_grob
logical. Should a snapshot of the display be
returned as a grid grob?
- a
intercept; alternatively, a regression model from which
coefficients can be extracted via coef
.
- b
slope.
- ...
Further arguments passed to grid.abline
.