Skip to contents

Compare MID Importances with ggplot2

Source: R/ggmid_midimps.R
ggmid.midimps.Rd

For "midimps" collection objects, ggmid() visualizes and compares the importance of component functions across multiple fitted MID models.

Usage

# S3 method for class 'midimps'
ggmid(
  object,
  type = c("barplot", "dotchart", "series"),
  theme = NULL,
  terms = NULL,
  max.nterms = 15L,
  labels = NULL,
  ...
)

# S3 method for class 'midimps'
autoplot(object, ...)

Arguments

object

a "midimps" collection object to be visualized.

type

the plotting style. One of "barplot", "dotchart", or "series".

theme

a character string or object defining the color theme. See color.theme for details.

terms

an optional character vector specifying which terms to display. If NULL, terms are automatically extracted from the object.

max.nterms

the maximum number of terms to display. Defaults to 15.

labels

an optional numeric or character vector to specify the model labels. Defaults to the labels found in the object.

...

optional parameters passed on to the main layer (e.g., geom_col).

Value

ggmid.midimps() returns a "ggplot" object.

Details

This is an S3 method for the ggmid() generic that creates a comparative importance plot from a "midimps" collection object. It visualizes the average contribution of component functions to the fitted MID models, allowing for easy comparison across different models.

The type argument controls the visualization style: The default, type = "barplot", creates a standard grouped bar plot where the length of each bar represents the overall importance of the term, positioned side-by-side by model label. The type = "dotchart" option creates a grouped dot plot, offering a clean alternative to the bar plot for visualizing and comparing term importance across models. The type = "series" option plots the importance trend over the models for each component function.

See also

ggmid.midimp, plot.midimps

Examples

data(mtcars, package = "datasets")

# Fit two different models for comparison
mid1 <- interpret(mpg ~ wt + hp + cyl, data = mtcars)
#> 'model' not passed: response variable in 'data' is used
mid2 <- interpret(mpg ~ (wt + hp + cyl)^2, data = mtcars)
#> 'model' not passed: response variable in 'data' is used

# Calculate importance for both models and combine them
imps <- midlist(
  "Main Effects" = mid.importance(mid1),
  "Interactions" = mid.importance(mid2)
)

# Create a comparative grouped bar plot (default)
ggmid(imps)


# Create a comparative dot chart with a specific theme
ggmid(rev(imps), type = "dotchart", theme = "Okabe-Ito")


# Create a series plot to observe trends across models
ggmid(imps, type = "series")