In luyouepiusf/ClusteringTree4: Supervised clustering with survival outcomes

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(ClusteringTree4)
library(dplyr)
library(grid)
library(gridtext)
library(survival)
set.seed(1984)

Let simulate a dataframe with 5 numeric column

nind<-500
ndim_numeric<-5
ndim_factor<-3
ft<-rexp(nind,3)
ct<-rexp(nind,3)
ot<-pmin(ft,ct)
delta<-ft<ct
xx_numeric<-matrix(rnorm(nind*ndim_numeric),nind,ndim_numeric)
colnames(xx_numeric)<-paste0("X",1:ncol(xx_numeric))

xx_factor<-matrix(NA,nind,ndim_factor)
xx_factor[,1]<-sample(c("AA","BB","CC","DD"),nind,replace=TRUE)
xx_factor[,2]<-sample(c("apple","banana","banana","cherry"),nind,replace=TRUE)
xx_factor[,3]<-sample(c("good","medium","poor"),nind,replace=TRUE)
colnames(xx_factor)<-paste0("V",1:ncol(xx_factor))

a_survival_tree<-survival_tree(
  time=ot,
  event=delta,
  matrix_numeric=xx_numeric,
  matrix_factor=xx_factor,
  significance=0.05,
  missing="weighted")

a_table<-tree_to_table(a_survival_tree$survival_tree)
a_table$survival[[2]]

plot_survival_tree(a_survival_tree)

Create a test datatest

nind_test<-200
set.seed(-1)
xx_numeric_test<-matrix(rnorm(nind_test*ndim_numeric),nind_test,ndim_numeric)
colnames(xx_numeric_test)<-paste0("X",1:ncol(xx_numeric_test))
xx_factor_test<-matrix(NA,nind_test,ndim_factor)
xx_factor_test[,1]<-sample(c("AA","BB","CC","DD"),nind_test,replace=TRUE)
xx_factor_test[,2]<-sample(c("apple","banana","banana","cherry"),nind_test,replace=TRUE)
xx_factor_test[,3]<-sample(c("good","medium","poor"),nind_test,replace=TRUE)
colnames(xx_factor_test)<-paste0("V",1:ncol(xx_factor_test))

weight<-predict_weight(a_survival_tree,xx_numeric_test,xx_factor_test)
rowSums(weight)

result<-predict_distance(a_survival_tree,xx_numeric_test,xx_factor_test)

heatmap(result$ind_distance)

luyouepiusf/ClusteringTree4 documentation built on Oct. 9, 2022, 9:06 p.m.