ggRandomForests (version 1.1.1)

airq_prtl: A randomForestSRC::plot.variable object.

Description

A cached object from randomForestSRC::plot.variable function for the New York Air Quality Measurements randomForestSRC regression forest airq_rf.

format

randomForestSRC::plot.variable object for regression

Details

For ggRandomForests examples and tests, as well as streamlining the R CMD CHECK for package release, we cache the computationally expensive operations from the randomForestSRC package.

We build a regression randomForest (airq_rf) with the airquality measurements data, then run the plot.variable function to generate the data for constructing partial dependence plots.

This "data set" is a cache of the plot.variable function, with partial=TRUE for the "Wind" variable. The data is then a risk adjusted variable dependence curve from the airq_rf random forest model.

The data were from New York, from May to September 1973. The data was obtained from the New York State Department of Conservation (ozone data) and the National Weather Service (meteorological data).

References

Chambers, J. M., Cleveland, W. S., Kleiner, B. and Tukey, P. A. (1983) Graphical Methods for Data Analysis. Belmont, CA: Wadsworth.

Ishwaran H. and Kogalur U.B. (2014). Random Forests for Survival, Regression and Classification (RF-SRC), R package version 1.5.4.

Ishwaran H. and Kogalur U.B. (2007). Random survival forests for R. R News 7(2), 25-31.

Ishwaran H., Kogalur U.B., Blackstone E.H. and Lauer M.S. (2008). Random survival forests. Ann. Appl. Statist. 2(3), 841-860.

See Also

airquality plot.variable rfsrc airq_rf gg_partial plot.gg_partial

Examples

## airq data
## Load the stored randomForestSRC object for airq
load(air1_rf, package="ggRandomForests)

# Save the "Wind" partial plot data
airq_prtl <- plot.variable(airq_rf, xvar.names = "Wind",
                          partial=TRUE, show.plots=FALSE)

# Generate a ggRandomForests plot.
gg_dta <- gg_partial(airq_prtl)
plot(gg_dta)