Dear list,
I have a spatial data frame made of 533 point observations. I'd to split it in
a
validation data set (133 points) and divide the remainig points in a number of
calibration data sets of 40 points each one(400/10)
The code I made works almost fine but there are some points that overlap: I
post
here an example using the meuse data set:
require(sp)
data(meuse)
coordinates(meuse)<- ~ x+y
set.seed(2540)
valid.id <- sample(1:dim(meuse@data)[1], 25)
meuse.val<- meuse[valid.id,]
str(meuse.val)
meuse.calibration<- meuse[!is.element(1:dim(meuse@data)[1], valid.id),]
str(meuse.calibration); dim(meuse.calibration)
cal.id<- vector("list",10)
set.seed(2540)
for(i in 1:10) {
cal.id[[i]]<- sample(1:dim(meuse.calibration@data)[1], 13)
meuse.calibration<-
meuse.calibration[!is.element(1:dim(meuse.calibration)[1], cal.id[1:i]),]
}
sapply(cal.id,length)
meuse.calibration<- meuse[!is.element(1:dim(meuse@data)[1], valid.id),]
#create again the calibration data set
dim(meuse.calibration@data)
meuse.cal<- vector("list",10)
for(i in 1:10) meuse.cal[[i]] <- meuse.calibration[cal.id[[i]],]
plot(meuse.cal[[1]])
for(i in 2:10) points(meuse.cal[[i]], pch=2:10, col=2:10)
points(meuse.val, pch=20, col="blue", cex=1.2)
rm(valid.id, cal.id)
Any help will be appreciate.
Thanks a lot
Giuseppe
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
