Visualize Stepwise Regression Results

Creates informative visualizations of the variable selection process from a StepReg object. This function generates two types of plots: detailed step-by-step selection process and an overview of the final selected variables.

# S3 method for class 'StepReg'
plot(
  x,
  strategy = attr(x, "nonhidden"),
  process = c("overview", "detail"),
  num_digits = 6,
  ...
)

Arguments

x

A StepReg object containing the results of stepwise regression analysis.

strategy

Character. Specifies which selection strategy to visualize:

• "forward" - Forward selection

• "backward" - Backward elimination

• "bidirection" - Bidirectional selection

• "subset" - Best subset selection

Default is the first strategy name in the StepReg object.

process

Character. Specifies the type of visualization to display:

• "detail" - Shows detailed step-by-step selection process with variable entry/removal

• "overview" - Shows summary of the selection process with metric values

Default is "overview".

num_digits

Integer. Number of decimal places to display in the plots. Default is 6.

...

Additional argument passed to plotting functions (currently not used).

Value

A ggplot object showing either:

• For "detail" process: A heatmap showing variable selection status at each step

• For "overview" process: A line plot showing metric values across steps

Details

The function creates different types of visualizations based on the selection strategy:

• For forward/backward/bidirectional selection:

  • detail view shows a heatmap with green tiles for added variables, tan tiles for removed variables, and gray tiles for non-selected variables

  • Overview shows metric values across steps with variable labels

• For subset selection:

  • detail view shows a heatmap of selected variables at each step

  • Overview shows metric values for different subset sizes

See also

stepwise for creating StepReg objects

Examples

if (FALSE) { # \dontrun{
# Load example data
data(mtcars)

# Run stepwise regression with multiple strategies
formula <- mpg ~ .
result <- stepwise(
  formula = formula,
  data = mtcars,
  type = "linear",
  strategy = c("forward", "bidirection", "subset"),
  metric = c("AIC", "BIC", "SL")
)

# Generate default overview plot
plot(result)

# Generate detailed plot for forward selection
plot(result, strategy = "forward", process = "detail")

# Generate overview plot with 3 decimal places
plot(result, strategy = "bidirection", process = "overview", num_digits = 3)
} # }