plotNull: Plot Random Forest Importance Null Distributions

In EricArcher/rfPermute: Estimate Permutation p-Values for Random Forest Importance Metrics

Plot Random Forest Importance Null Distributions

Description

Plot the Random Forest null distributions importance metrics, observed values, and p-values for each predictor variable from the object produced by a call to rfPermute.

Usage

plotNull(
  x,
  preds = NULL,
  imp.type = NULL,
  scale = TRUE,
  plot.type = c("density", "hist"),
  plot = TRUE
)

Arguments

x	An object produced by a call to rfPermute.
preds	a character vector of predictors to plot. If NULL, then all predictors are plotted.
imp.type	A character vector giving the importance metric(s) to plot.
scale	Plot importance measures scaled (divided by) standard errors?
plot.type	type of plot to produce: "density" for smoothed density plot, or "hist" for histogram.
plot	display the plot?

Details

The function will generate an plot for each predictor, with facetted importance metrics. The vertical red line shows the observed importance score and the _p_-value is given in the facet label.

Value

A named list of the ggplot figures produced is invisibly returned.

Author(s)

Eric Archer [e-mail]

Examples

# A regression model using the ozone example
data(airquality)
ozone.rp <- rfPermute(
  Ozone ~ ., data = airquality, ntree = 100, 
  na.action = na.omit, nrep = 50, num.cores = 1
)
  
# Plot the null distributions and observed values.
plotNull(ozone.rp) 

EricArcher/rfPermute documentation built on Aug. 28, 2023, 2:37 a.m.