plotInteraction: Plots response surfaces for tree-based models.

In memoria: Quantifying Ecological Memory in Palaeoecological Datasets and Other Long Time-Series

Description

Plots a response surface plot or interaction plot (2 predictors and a model response) for models of the functions ranger, randomForest, and rpart. It also plots the observed data on top of the predicted surface.

Usage

plotInteraction(
  model = NULL,
  data = NULL,
  x = NULL,
  y = NULL,
  z = NULL,
  grid = 100,
  point.size.range = c(0.1, 1)
  )

Arguments

model	a model object produced by the functions ranger, randomForest, or rpart.
data	dataframe used to fit the model.
x	character string, name of column in data to be plotted in the x axis.
y	character string, name of column in data to be plotted in the y axis.
z	character string, name of column in data to be plotted as a surface. Generally, the name of the response variable in model.
grid	numeric, resolution of the x and y axes.
point.size.range	numeric vector with two values defining the range size of the points representing the observed data.

Value

A ggplot object.

Author(s)

Blas M. Benito <[e-mail]>

memoria documentation built on May 17, 2019, 9 a.m.