getTimeROC: time-dependent ROC curve base on timeROC package

In shijianasdf/BasicBioinformaticsAnalysisFromZhongShan: Launch Usual Clinical tool for Kind Yield

Description

getTimeROC is a powerful function to do time-dependent ROC curve base on timeROC::timeROC.getTimeROC help quickly and conveniently set common parameter in a ROC curve of ggplot2 style.

Usage

getTimeROC(data, time.col, status.col, cause, markers,
  merge.markers = NULL, weighting.univariate = c("marginal", "cox",
  "aalen")[2], weighting.multivariate = c("marginal", "cox", "aalen")[2],
  time.raw.type = c("Day", "Month", "Year")[1],
  time.target.type = c("Day", "Month", "Year")[3], time.knot = c(3, 5),
  line.color = mycolor, size = 1.2, half.border = F,
  legend.position = c(0.8, 0.2), show.auc = T, auc.digits = 2,
  reference = T, width = 10, height = 10, merge.plot = T,
  names = "love")

Arguments

data	a data frame
time.col	time colnames
status.col	status
cause	the event status.For example,if 1=death,0=alive,then cause should be 1.
markers	The vector of the marker values for which we want to compute the time-dependent ROC curves. Without loss of generality, the function assumes that larger values of the marker are associated with higher risks of events. If lower values of the marker are associated with higher risks of events, then reverse the association adding a minus to the marker values.
merge.markers	a list of markers groups.Default is NULL.
weighting.univariate	Default is "marginal"
weighting.multivariate	Default is "cox"
time.raw.type	one of "Day","Month","Year"
time.target.type	one of "Day","Month","Year"
time.knot	target time knots.Like 3 years and 5 years survival.
line.color	line colors in ROC plot.
size	the size of plot
half.border	whether show half border of the plot
legend.position	the position of legend position
width	the width of plot
height	the height of plot
merge.plot	whether use a merge style by time.knot
names	part of output file name.

Value

a list contain model information and series ggplot object.

Author(s)

Weibin Huang<[e-mail]>

See Also

timeROC.

Examples

library(lucky)
library(survival)
data(pbc)
pbc<-pbc[!is.na(pbc$trt),]
pbc$status<-as.numeric(pbc$status==2)
data = pbc
time.col = "time"
status.col ="status"
cause = 1
markers = c("bili","chol")
merge.markers=list(c("bili","chol"))
time.raw.type=c("Day","Month","Year")[1]
time.target.type=c("Day","Month","Year")[3]
time.knot=c(3,5)
names = "love"
line.color = mycolor[c(4,5,3,8,23)]
legend.position = "top"
weighting.univariate = "marginal"
weighting.multivariate = "cox"
size = 1.2
half.border=T
reference = F
width = 10
height=10
merge.plot = T
names = "love"

## merge style
models.merge <- getTimeROC(data,
                           time.col,status.col,cause,
                           markers,
                           merge.markers,
                           time.raw.type=time.raw.type,
                           time.target.type=time.target.type,
                           time.knot=time.knot,
                           line.color = line.color,
                           half.border=half.border,
                           reference = reference,
                           names = "love")

## single style
models.single <- getTimeROC(data,
                            time.col,status.col,cause,
                            markers,
                            merge.markers,
                            time.raw.type=time.raw.type,
                            time.target.type=time.target.type,
                            time.knot=time.knot,
                            line.color = line.color,
                            half.border=half.border,
                            reference = reference,
                            merge.plot = F,
                            names = "love")

## Paquid
library(timeROC);library(survival)
data(Paquid) # rm(Paquid)
Paquid$DSST <- -Paquid$DSST
Paquid$status <- ifelse(Paquid$status == 1,1,0)
models.merge <- getTimeROC(data = Paquid,
                           time.col = "time",
                           status.col = "status",
                           cause = 1,
                           markers = "DSST",
                           merge.markers = list(
                             merge = c("DSST","MMSE")
                           ),
                           time.raw.type="Year",
                           time.target.type="Year",
                           time.knot= c(3,5),
                           line.color = mycolor[c(1,3,4)],
                           names = "Paquid1")

shijianasdf/BasicBioinformaticsAnalysisFromZhongShan documentation built on Jan. 3, 2020, 10:08 p.m.