pdpZenplot: Create a zenplot displaying partial dependence values.
In AlanInglis/vividOld: Variable Importance and Variable Interaction Displays

Description Usage Arguments Value Examples

View source: R/zPlot.R

Constructs a zigzag expanded navigation plot (zenplot) displaying partial dependence values.

pdpZenplot(
  task,
  model,
  zpath = NULL,
  method = "pdp",
  noCols = c("letter", "square", "A4", "golden", "legal"),
  zenMethod = c("tidy", "double.zigzag", "single.zigzag", "rectangular"),
  pal = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
  fitlims = "pdp",
  gridSize = 10,
  nmax = 500,
  class = 1,
  ...
)

`task`	Task created from the mlr3 package, either regression or classification.
`model`	A machine learning model created from mlr3 task and learner.
`zpath`	A zenpath created from calcZpath. see `zenpath` from the `zenplots` package for more details.
`method`	"pdp" (default) or "ale"
`noCols`	number of columns of 2d plots (>= 1) or one of "letter", "square", "A4", "golden" or "legal" in which case a similar layout is constructed. See ?zenplot
`zenMethod`	String indicating the layout of the zigzag plot. The available methods are: "tidy": more tidied-up double.zigzag (slightly more compact placement of plots towards the end). "double.zigzag": zigzag plot in the form of a flipped “S”. Along this path, the plots are placed in the form of an “S” which is rotated counterclockwise by 90 degrees. "single.zigzag": zigzag plot in the form of a flipped “S”. "rectangular": plots that fill the page from left to right and top to bottom. This is useful (and most compact) for plots that do not share an axis.
`pal`	A vector of colors to show predictions, for use with scale_fill_gradientn
`fitlims`	Specifies the fit range for the color map. Options are a numeric vector of length 2, "pdp" (default), in which cases limits are calculated from the pdp. Predictions outside fitlims are squished on the color scale.
`gridSize`	for the pdp/ale plots, defaults to 10.
`nmax`	Maximum number of data rows to consider, for calculating pdp.
`class`	For a classification model, show the probability of this class. Defaults to 1.

A zenplot of partial dependence values.

# Load in the data:
aq <- na.omit(airquality)*1.0

# Run an mlr3 ranger model:
library(mlr3)
library(mlr3learners)
library(ranger)
ozonet  <- TaskRegr$new(id = "airQ", backend = aq, target = "Ozone")
ozonel  <- lrn("regr.ranger", importance = "permutation")
ozonef  <- ozonel$train(ozonet)

# Create matrix
viv <- vividMatrix(ozonet, ozonef)

# Calculate Zpath:
zpath<-calcZpath(viv,.8)
zpath

# Create graph:
pdpZenplot(ozonet, ozonef, zpath=zpath)