Inverse Distance Weighted smoothing
Distance weighted smoothing of a variable in a spatial point object
idw.smoothing(x, y, d, k)
x |
Object of class SpatialPointsDataFrame |
y |
Numeric data in x@data |
d |
Distance constraint |
k |
Maximum number of k-nearest neighbors within d |
A vector, same length as nrow(x), of adjusted y values
Smoothing is conducted with a weighted-mean where; weights represent inverse standardized distance lags Distance-based or neighbour-based smoothing can be specified by setting the desired neighbour smoothing method to a specified value then the other parameter to the potential maximum. For example; a constraint distance, including all neighbors within 1000 (d=1000) would require k to equal all of the potential neighbors (n-1 or k=nrow(x)-1).
library(sp)
data(meuse)
coordinates(meuse) <- ~x+y
# Calculate distance weighted mean on cadmium variable in meuse data
cadmium.idw <- idw.smoothing(meuse, 'cadmium', k=nrow(meuse), d = 1000)
meuse@data$cadmium.wm <- cadmium.idw
opar <- par(no.readonly=TRUE)
par(mfrow=c(2,1))
plot(density(meuse@data$cadmium), main='Cadmium')
plot(density(meuse@data$cadmium.wm), main='IDW Cadmium')
par(opar)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.