R/09.btrm_print.R

In btrm: Bayesian Treed Regression Model for Personalized Prediction and Precision Diagnostics

btrm_print=function(ET){
  x=ET$x
  z=ET$z

  internal=ET$internal
  terminal=ET$terminal
  cutoff=ET$cutoff
  splitVariable=ET$splitVariable
  beta.hat=ET$beta.hat

  colX=colnames(x)
  if(is.null(colX[1])){
    Q=ET$q
    colX=paste0("x",1:Q)
  }

  colZ=colnames(z)
  if(is.null(colZ[1])){
    P=ET$p
    colZ=paste0("Z",1:P)
  }

  if(terminal[1]==1){
    message(paste0("Root node (or node 1) - ML model: ",ET$selML[1]))
  }else{
    for(s in terminal){
      depth=floor(log2(s))
      N1=paste0(rep("-",depth),collapse="")
      N2=paste0("Node ",s,collapse="")
      N3=paste0(colZ[ET$marker[s]]," (b0=",beta.hat[[s]][1],", b1=",beta.hat[[s]][2],")")

      N4=list()
      depth2=depth
      s1=s #terminal node
      for(j in depth:1){
        s2=floor(s1/2) #internal node above s1
        dir=">" #right
        if(s1%%2==0) #left
          dir="<="
        N4[[j]]=paste0(colX[splitVariable[s2]]," ",dir," ",cutoff[s2])
        s1=s2
      }
      if(depth>=2)
        N4=paste0(unlist(N4),collapse=" & ")

      N=paste0(c(N1,N2,": (",N4,"), ",N3),collapse="")
      message(N)
    }
  }
}