Skip to contents

Plot method for checking model assumptions

Source: R/plot.check_model.R
plot.see_check_model.Rd

The plot() method for the performance::check_model() function. Diagnostic plots for regression models.

Usage

# S3 method for class 'see_check_model'
plot(
  x,
  theme = NULL,
  colors = NULL,
  type = c("density", "discrete_dots", "discrete_interval", "discrete_both"),
  n_columns = 2,
  show_smooth = NULL,
  ...
)

Arguments

x

An object.

theme

A ggplot2-theme function, e.g. theme = theme_lucid() or theme = ggplot2::theme_dark().

colors

Character vector of length two, indicating the colors (in hex-format) for points and line.

type

Plot type for the posterior predictive checks plot. Can be "density" (default), "discrete_dots", "discrete_interval" or "discrete_both" (the discrete_* options are appropriate for models with discrete - binary, integer or ordinal etc. - outcomes).

n_columns

Number of columns to align plots.

show_smooth

Logical, if FALSE, no trend line is drawn in the Linearity, Homogeneity of Variance and Influential Observations panels. This turns off the LOESS smoother and, for models whose fitted values take only a few distinct values, the cell means with their line and error bars. If NULL (default), the value stored in the check_model() object is used, and TRUE if there is none. The binned residuals panel is not affected.

...

Arguments passed to or from other methods.

Value

A ggplot2-object.

Details

Larger models (with many observations) may take a longer time to render. Thus, the number of data points is limited to 2000 by default. Use plot(check_model(), maximum_dots = <number>) (or check_model(maximum_dots = <number>)) to define the number of data points that should be shown in the plots.

See also

See also the vignette about check_model().

Examples

if (FALSE) { # require("patchwork") && FALSE
library(performance)

model <- lm(qsec ~ drat + wt, data = mtcars)
plot(check_model(model))
}