ggRandomForests (version 1.1.1)

pbc_ggpart: A gg_partial object for the pbc data, combined from 1 and 3 year plot.variable for pbc data.

Description

This object is built for use in the vignette "Survival with Random Forests".

format

A gg_partial object for survival

References

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.

Examples

data(pbc_rf)
data(pbc_vs)

# For the top 6 covariates of the rfsrc object by minimal depth
xvar <- pbc_vs$topvars[1:6]

# Calculate the 1 year partial dependence
pbc_prtl <- plot.variable(pbc_rf, surv.type="surv",
                         time=1,
                         xvar.names=xvar, partial=TRUE,
                         show.plots = FALSE)


# Calculate the 3 year partial dependence
pbc_prtl.3 <- plot.variable(pbc_rf, surv.type="surv",
                            time=3,
                            xvar.names=xvar, partial=TRUE,
                            show.plots = FALSE)

# Create gg_partial objects
gg_dta <- gg_partial(pbc_prtl)
gg_dta.3 <- gg_partial(pbc_prtl.3)

# Combine the objects to get multiple time curves
# along variables on a single figure.
pbc_ggpart <- combine(gg_dta, gg_dta.3,
                      labels=c("1 Year", "3 Years"))