ggInfluence: ggInfluence

In JBjouffray/ggBRT: Explore and Visualise the Results of Boosted Regression Trees

Description

Function to return and plot the relative influence of predictors from boosted regression trees obtained with the gbm.step routine in the dismo package.

Usage

ggInfluence(gbm.object, col.bar = "skyblue3", show.signif = TRUE,
  col.signif = "#8B3A3A", main = gbm.call$response.name,
  x.label = "Relative influence (%)", signif = FALSE, plot = T, ...)

Arguments

gbm.object	a gbm.step object (object of S3 class gbm)
col.bar	color of the bars
show.signif	Logical. If TRUE (default), display a vertical dashed line indicating the threshold of significance (100/number of predictors, Muller et al. 2013)
col.signif	color of the dashed line
main	title for the plot (by default the name of the response variable)
x.label	title for the x-axis
signif	Logical. If TRUE (default FALSE), retains only the "significant" predictors (Muller et al. 2013) and rescale their influence to 100
plot	Logical. If TRUE (default) draws a horizontal barplot of the relative influence of each predictor

Value

Returns a dataframe with the relative influence of each predictor in the model and a barplot

JBjouffray/ggBRT documentation built on Jan. 30, 2021, 8:13 a.m.