permVimp: permVimp

In AlanInglis/BartVis: Create Visualisations for BART Models

permVimp

Description

A variable selection approach which creates a null model by permuting the response, rebuilding the model, and calculating the inclusion proportion (IP) on the null model. The final result displayed is the original model's IP minus the null IP.

Usage

permVimp(model, data, response, numTreesPerm = NULL, plotType = "barplot")

Arguments

model	Model created from either the BART, dbarts or bartMachine packages.
data	A data frame containing variables in the model.
response	The name of the response for the fit.
numTreesPerm	The number of trees to be used in the null model. As suggested by Chipman (2009), a small number of trees is recommended (~20) to force important variables to used in the model. If NULL, then the number of trees from the true model is used.
plotType	Either a bar plot ('barplot') or a point plot ('point')

Value

A variable selection plot.

Examples

if (requireNamespace("dbarts", quietly = TRUE)) {
 # Load the dbarts package to access the bart function
 library(dbarts)
 # Get Data
 df <- na.omit(airquality)
 # Create Simple dbarts Model For Regression:
 set.seed(1701)
 dbartModel <- bart(df[2:6],
   df[, 1],
   ntree = 5,
   keeptrees = TRUE,
   nskip = 10,
   ndpost = 10
 )
 # Tree Data
 trees_data <- extractTreeData(model = dbartModel, data = df)
 permVimp(model = dbartModel, data = df, response = 'Ozone', numTreesPerm = 2, plotType = 'point')
}

AlanInglis/BartVis documentation built on July 27, 2024, 12:02 a.m.