gg_partial: Partial variable dependence object

Description

The randomForestSRC::plot.variable function returns a list of either marginal variable dependance or partial variable dependence data from a randomForestSRC::rfsrc object. The gg_partial function formulates the randomForestSRC::plot.variable output for partial plots (where partial=TRUE) into a data object for creation of partial dependence plots using the plot.gg_partial function.

Partial variable dependence plots are the risk adjusted estimates of the specified response as a function of a single covariate, possibly subsetted on other covariates.

Usage

gg_partial.rfsrc(object, named, ...)

Arguments

object

the partial variable dependence data object from randomForestSRC::plot.variable function

named

optional column for merging multiple plots together

...

optional arguments

Value

gg_partial object. A data.frame or list of data.frames corresponding the variables contained within the randomForestSRC::plot.variable output.

References

Friedman, Jerome H. 2000. "Greedy Function Approximation: A Gradient Boosting Machine." Annals of Statistics 29: 1189-1232.

Examples

Run this code

## ------------------------------------------------------------
## classification
## ------------------------------------------------------------
## -------- iris data

## iris "Petal.Width" partial dependence plot
##
# rfsrc_iris <- rfsrc(Species ~., data = iris)
# partial_iris <- plot.variable(rfsrc_iris, xvar.names = "Petal.Width",
#                            partial=TRUE)
data(partial_iris, package="ggRandomForests")

gg_dta <- gg_partial(partial_iris)
plot(gg_dta)

## ------------------------------------------------------------
## regression
## ------------------------------------------------------------
## -------- air quality data
## airquality "Wind" partial dependence plot
##
# rfsrc_airq <- rfsrc(Ozone ~ ., data = airquality)
# partial_airq <- plot.variable(rfsrc_airq, xvar.names = "Wind",
#                            partial=TRUE, show.plot=FALSE)
data(partial_airq, package="ggRandomForests")

gg_dta <- gg_partial(partial_airq)
plot(gg_dta)

gg_dta.m <- gg_dta[["Month"]]
plot(gg_dta.m, notch=TRUE)

gg_dta[["Month"]] <- NULL
plot(gg_dta, panel=TRUE)

## -------- Boston data
data(partial_Boston, package="ggRandomForests")

gg_dta <- gg_partial(partial_Boston)
plot(gg_dta, panel=TRUE)

## -------- mtcars data
data(partial_mtcars, package="ggRandomForests")
gg_dta <- gg_partial(partial_mtcars)

gg_dta.cat <- gg_dta
gg_dta.cat[["disp"]] <- gg_dta.cat[["wt"]] <- gg_dta.cat[["hp"]] <- NULL
gg_dta.cat[["drat"]] <- gg_dta.cat[["carb"]] <- gg_dta.cat[["qsec"]] <- NULL

plot(gg_dta.cat, panel=TRUE, notch=TRUE)

gg_dta[["cyl"]] <- gg_dta[["vs"]] <- gg_dta[["am"]] <- NULL
gg_dta[["gear"]] <- NULL
plot(gg_dta, panel=TRUE)
## ------------------------------------------------------------
## survival examples
## ------------------------------------------------------------
## -------- veteran data
## survival "age" partial variable dependence plot
##
# data(veteran, package = "randomForestSRC")
# rfsrc_veteran <- rfsrc(Surv(time,status)~., veteran, nsplit = 10, ntree = 100)
#
## 30 day partial plot for age
# partial_veteran <- plot.variable(rfsrc_veteran, surv.type = "surv",
#                               partial = TRUE, time=30,
#                               xvar.names = "age",
#                               show.plots=FALSE)
data(partial_veteran, package="ggRandomForests")

gg_dta <- gg_partial(partial_veteran[[1]])
plot(gg_dta)

gg_dta.cat <- gg_dta
gg_dta[["celltype"]] <- gg_dta[["trt"]] <- gg_dta[["prior"]] <- NULL
plot(gg_dta, panel=TRUE)

gg_dta.cat[["karno"]] <- gg_dta.cat[["diagtime"]] <- gg_dta.cat[["age"]] <- NULL
plot(gg_dta.cat, panel=TRUE, notch=TRUE)

gg_dta <- lapply(partial_veteran, gg_partial)
gg_dta <- combine.gg_partial(gg_dta[[1]], gg_dta[[2]] )

plot(gg_dta[["karno"]])
plot(gg_dta[["celltype"]])

gg_dta.cat <- gg_dta
gg_dta[["celltype"]] <- gg_dta[["trt"]] <- gg_dta[["prior"]] <- NULL
plot(gg_dta, panel=TRUE)

gg_dta.cat[["karno"]] <- gg_dta.cat[["diagtime"]] <- gg_dta.cat[["age"]] <- NULL
plot(gg_dta.cat, panel=TRUE, notch=TRUE)
## -------- pbc data