build_MOTTE_tree_CO: Fitting MOTTE Tree in the MOTTE.RF

In boyiguo1/MOTTE.RF: What the Package Does (One Line, Title Case)

Description

Fitting MOTTE Tree in the MOTTE.RF

Usage

build_MOTTE_tree_CO(
  x.b,
  x.diff.1,
  x.diff.2,
  y.diff.1,
  y.diff.2,
  nodesize,
  nsplits,
  left.out
)

Arguments

x.b	Pre-treatment covariates, i.e. microbiomes
nodesize	An integer value. The threshold that control the maximum size of a node
nsplits	A numeric value, the number of maximum splits
left.out	left.out is ensure at least left.out*2 sample for either treated or untreated sample in the group
x.e	Post-treatment covariates, i.e. microbiomes
treat	A vector of binary value, the arm of treatment
y.b	Pre-treatment response, i.e. biomarkers
y.e	Post-treatment response, i.e. biomarkers
seed	a seed number to generate the random subsets of split candidates. Doesn't work when applying nsplit==NULL

Value

A data.tree object, node

Examples

set.seed(1)
B <- create.B(10)
Z <- create.Z(10, 3)

tmp.dat <- sim_MOTTE_data( n.train = 500, n.test = 200,
p = 10, q = 3, ratio = 0.5,
B = B, Z = Z)

train.dat <- tmp.dat$train

x.b <- scale(train.dat$x.b, center = FALSE, scale = TRUE)
x.e <- scale(train.dat$x.e, center = FALSE, scale = TRUE)
y.b <- scale(train.dat$y.b, center = FALSE, scale = TRUE)
y.e <- scale(train.dat$y.e, center = FALSE, scale = TRUE)
treat <- train.dat$trt
#with(train.dat,
    build_MOTTE_tree(x.b, x.e, factor(treat), y.b, y.e,
                     nodesize=30, nsplits=NULL, left.out = 0.1)
 #)

boyiguo1/MOTTE.RF documentation built on June 14, 2020, 4:12 p.m.