gw_do_weight: Application of the geographic weighting function

In gwverse/gw: GW Core Package

Application of the geographic weighting function

Description

This function is for use within the application related function, for example to undertake a GWR. The idea here is that rather than working with a single observation location, the analysis is procededing with a vector of locations and multiple indices of nearby locations in matrix form for adaptive bandwidths (ie the nearest *n* observation) or in list form for fixed bandwidths (ie the observations within a specified distance of the different locations, whose number may vary from location to location). This assembles the outputs of 'gw_get_weight' and 'gw_get_nearby' so that they can be used in an 'apply' function across the whole spatial dataset rather than for just a single observation at a time.

Usage

gw_do_weight(i, bw, nearbyMat, dist_mat, weight_func)

Arguments

i	The i^th observation (as 'obs_index' in 'gw_get_weight' and 'gw_get_nearby')
bw	The bandwidth (fixed or adaptive).
nearbyMat	A 'matrix' or 'list' of indices of nearby locations for each observation.
dist_mat	An n * n matrix of distances of each location.
weight_func	The weighting function returned by 'gw_get_weight'.

Value

A vector of weights, for use in for example a locally weighted regression model.

Examples

# load some packages and data
library(tmap)
data(georgia)
# define a distance matrix, a location and an adaptive bandwidth
dist_mat = as.matrix(dist(st_coordinates(st_centroid(georgia)), upper = T, diag = T))
# create the nearby function - see the help for `gw_get_nearby`
nearby_func = gw_get_nearby(adaptive = TRUE)
# create the weighting function - see the help for `gw_get_weight`
weight_func = gw_get_weight(kernel = "bisquare", adaptive = TRUE)
# create an index of observations
indexMat  = matrix(1:nrow(georgia), ncol = 1) 	
# determine nearby locations for all observation (list for adaptive, matrix for fixed bandwidth) 
nearbyMat = apply(indexMat, 1, function(x) 
  nearby_func(x, dist_mat, bw))
# apply the weight function to the nearby locations 
weightMat = apply(indexMat, 1, function(x) 
  gw_do_weight(x, bw, nearbyMat, dist_mat, weight_func))
# examine
dim(weightMat)
weightMat[, 70]
# map the result
georgia$weight = weightMat[, 70]
tm_shape(georgia)+tm_polygons("weight", palette = "Reds")+
tm_shape(georgia[70,])+tm_borders(lwd = 2) 

gwverse/gw documentation built on May 4, 2022, 1:46 a.m.