PartitionParty: Plot the partitions of a partykit tree model

In TreeHotspots: Hotspot Detection using Classification Trees

Description

COmputes and plots the partitions of a partykit tree model in 2 dimensions.

Usage

PartitionParty(ct, vars = c("lon", "lat"), BBOX, PLOT = TRUE, 


    labs = TRUE, verbose = 0)

Arguments

ct	object of class partykit
vars	variable names in 2D
BBOX	bounding box
PLOT	overlay retangles?
labs	add node ID labels to plot ?
verbose	level of verbosity

Author(s)

Markus Loecher

Examples

if (0){


  


  library(partykit)


  library(rpart)


  #unimodal bump


  data(data1, envir = environment())


  rp = rpart(l ~ x + y, data = data1)


  #plot(rp)


  party_rp <- partykit::as.party(rp)


  plot(party_rp)


  


  plot(y ~ x, col = as.numeric(l), data = data1, cex = 0.5, pch=20)


  PartitionParty(party_rp,vars=c("x","y"), verbose=0)


  


  #elliptical bump:


  data(data2, envir = environment())


  rp = rpart(l ~ x + y, data = data2)


  #plot(rp)


  party_rp <- as.party(rp)


  plot(party_rp)


  


  plot(y ~ x, col = as.numeric(l)-1, data = data2, cex = 0.5, pch=20)


  PartitionParty(party_rp,vars=c("x","y"), verbose=0) 


  


  #Artifial data at multiple scales and angles:


  data(msdata, envir = environment())


  ct <- ctree(clus ~ lon+lat,data = msdata,


              control = ctree_control(multiway=TRUE))


  


  plot(ct)


  plot(lat ~ lon,data = msdata, col = clus, pch=20,cex=0.6)


  PartitionParty(ct,vars=c("lon","lat"), verbose=0)


}


print(1)

TreeHotspots documentation built on May 2, 2019, 5:17 p.m.