correctionTree: correctionTree - builds a binary tree of small zone...

Description Usage Arguments Details Value Examples

Description

correctionTree - builds a binary tree of small zone corrections

Usage

1
2
3
4
correctionTree(qProb, map, pErr = 0.9, optiCrit = 2, minSize = 0.012,
  minSizeNG = 0.001, distIsoZ = 0.075, simplitol = 0.001, LEQ = 5,
  MAXP = 0.1, LASTPASS = TRUE, disp = 0, SAVE = TRUE, ONE = FALSE,
  ALL = FALSE)

Arguments

qProb

probability vector used to generate quantile values

map

object returned by function genMap

pErr

equality tolerance for distance calculations

optiCrit

criterion choice

minSize

zone area threshold under which a zone is too small to be manageable

minSizeNG

zone area threshold under which a zone will be removed

distIsoZ

threshold distance to next zone, above which a zone is considered to be isolated

simplitol

tolerance for spatial polygons geometry simplification

LEQ

length of quantile sequence used to grow isolated zone

MAXP

quantile sequence maximum shift quantile sequence maximum shift

LASTPASS

if TRUE, remove zones that are still too small at the last level of the correction tree

disp

0: no info, 1: some info, 2: detailed info

SAVE

logical value, if TRUE function returns last level zonings, if FALSE function only returns best last level results

ONE

logical value, if TRUE function returns only criterion value

ALL

logical value, if TRUE function returns zonings at all levels

Details

builds a binary tree of possible corrections for small zone removal in a zoning. The zoning is based on contour lines corresponding to quantile values. These quantiles correspond the given probabilities. At each level, 2 branches (at max) are built, one for small zone removal, one for small zone growing or junction to another one. If growing or junction is not valid, the branch is not developped. Growing is done if the zone is isolated from others (see distIsoZ argument and optiGrow function). Junction is done if the zone is not isolated, and if there is a zone close by having the same label (see optirG function). At the last level, all zones will have been corrected, and the resulting zonings are evaluated regarding criteria and costs.

Value

a list with components

bestcrit

best criterion value at last level (in all cases)

critList

criterion values at last level (in all cases if ONE=FALSE)

costList

cost values at last level (in all cases if ONE=FALSE)

costLList

cost per label values at last level (in all cases if ONE=FALSE)

nzList

vector of number of zones at last level (in all cases if ONE=FALSE)

qProb

vector of probabilities values used for quantiles (in all cases if ONE=FALSE)

zk

list of zoning objects (such as returned by calNei function), first element corresponds to initial zoning, each other element is a list with each (last if ALL=FALSE) level zoning objects (only if SAVE=TRUE)

mdist

list of initial distance matrix and all (last if ALL=FALSE) level distance matrices (only if SAVE=TRUE)

criterion

list of initial criterion and all (last if ALL=FALSE) level criteria (only if SAVE=TRUE)

cost

list of initial cost and all (last if ALL=FALSE) level costs (only if SAVE=TRUE)

costL

list of initial cost per label and all (last if ALL=FALSE) level costs per label (only if SAVE=TRUE)

nz

list of initial number of zones and all (last if ALL=FALSE) level number of zones (only if SAVE=TRUE)

Examples

1
2
3
4
data(mapTest)
criti=correctionTree(c(0.4,0.7),mapTest,SAVE=TRUE) 
plotZ(criti$zk[[1]][[1]]$zonePolygone)
plotZ(criti$zk[[2]][[1]]$zonePolygone) # zones 7 and 8 were handled

geozoning documentation built on May 2, 2019, 9:43 a.m.