uni.tree {uni.survival.tree}

A survival tree based on stabilized score tests

Description

This function returns a classification (decision) tree for a given survival dataset. The decision of making inner nodes (splitting criterion) is based on the univariate score tests. The decision of declaring terminal nodes (stopping criterion) is the P-value threshold given by an argument. This tree construction algorithm is proposed by Emura et al. (2021).

Usage

uni.tree(
  t.vec,
  d.vec,
  X.mat,
  P.value = 0.01,
  d0 = 0.01,
  S.plot = FALSE,
  score = TRUE
)

Arguments

t.vec

:Vector of survival times (time to either death or censoring)

d.vec

:Vector of censoring indicators (1=death, 0=censoring)

X.mat

:n by p matrix of covariates (features), where n is the sample size and p is the number of covariates

P.value

:the threshold of P-value for stop splitting (stopping criterion)

d0

:A positive constant to stabilize the variance of score statistics (Witten & Tibshirani 2010)

S.plot

:call for plot the KM estimator for each split

score

:TRUE = score test (Emura et al. 2019); FALSE = log-rank test

Details

In order to stabilize the univariate score tests, a small value "d0" is added to the variance of the score statistics (Witten and Tibshirani 2010). d0=0 corresponds to the logrank test. To perform a large number of the score tests, the "compound.Cox" packages (Emura et al.2019) is applied with d0 as a option.

Value

A nested list describing a classification tree, consisting of inner nodes and terminal node.

References

Emura T, Hsu WC, Chou WC (2021). A survival tree based on stabilized score tests for high-dimensional covariates, in review

Emura T, Matsui S, Chen HY (2019). compound.Cox: Univariate Feature Selection and Compound Covariate for Predicting Survival, Computer Methods and Programs in Biomedicine 168: 21-37.

Witten DM, Tibshirani R (2010) Survival analysis with high-dimensional covariates. Stat Method Med Res 19:29-51

Examples

data(Lung,package="compound.Cox")
train_Lung=Lung[which(Lung[,"train"]==TRUE),] #select training data
t.vec=train_Lung[,1]
d.vec=train_Lung[,2]
x.mat=train_Lung[,-c(1,2,3)]
uni.tree(t.vec,d.vec,x.mat,P.value=0.01,d0=0.01,S.plot=FALSE,score=TRUE)