optRF (version 1.2.0)

plot_stability: Plot random forest stability

Description

Plot the estimated stability of random forest against certain numbers of trees

Usage

plot_stability(
  optRF_object,
  measure = c("selection", "importance", "prediction"),
  from = 0,
  to = 1e+05,
  add_recommendation = TRUE,
  add = FALSE,
  ...
)

Value

A plot showing the estimated stability of random forest for the given num.trees values.

Arguments

optRF_object

An optRF_object, either the result from the opt_importance or the opt_prediction function.

measure

A character string indicating which stability measure is to be plotted. One of "selection" (default, visualises selection stability), "prediction" (visualises prediction stability) or "importance" (visualises variable importance stability).

from

Smallest num.trees value to be plotted.

to

Greatest num.trees value to be plotted.

add_recommendation

When set as TRUE, if a recommendation was stated within the opt_prediction or opt_importance function, the recommended num.trees value as well as the expected random forest stability will be highlighted in the graph

add

If FALSE, a new plot will be created, if TRUE, the graph will be added to an existing plot.

...

Any other arguments from the plot function.

Examples

if (FALSE) {
data(SNPdata)
set.seed(123)
result_optpred = opt_prediction(y = SNPdata[,1], X=SNPdata[,-1]) # optimise random forest
plot_stability(result_optpred, measure = "prediction", add_recommendation = TRUE, add=FALSE)
plot_stability(result_optpred, measure = "selection",  add_recommendation = FALSE, add=TRUE)
}