
find an evenly spaced grid for x
makeraster(x, startx, d, newlevels, tol = 0.1)fitraster(x, tol = 0.1)
numeric to be fitted with a raster
starting point ("origin") for calculation of the raster
step size of the raster
levels of the raster
tolerance for rounding to new levels: elements of x within tol
of the distance between the levels of the new grid are rounded to the new grid point.
list with elements
the values of x
, possibly rounded to the raster values
the values of the raster
makeraster
fits the data to the specified raster.
fitraster
tries different raster parameter and returns the raster that covers most of the
x
values: The differences between the values of x
are calculated (possible step
sizes). For each of those step sizes, different points are tried (until all points have been
covered by a raster) and the parameter combination leading to the best coverage (i.e. most points
on the grid) ist used.
Note that only differences between the sorted values of x are considered as step size.
# NOT RUN {
x <- c (sample (1:20, 10), (0 : 5) + 0.5)
raster <- makeraster (x, x [1], 2)
raster
plot (x)
abline (h = raster$levels, col = "#00000040")
## unoccupied levels
missing <- setdiff (raster$levels, raster$x)
abline (h = missing, col = "red")
## points acutally on the raster
onraster <- raster$x %in% raster$levels
points (which (onraster), raster$x [onraster], col = "blue", pch = 20)
raster <- fitraster (x)
raster
plot (x)
abline (h = raster$levels, col = "#00000040")
## unoccupied levels
missing <- setdiff (raster$levels, raster$x)
abline (h = missing, col = "red")
## points acutally on the raster
onraster <- raster$x %in% raster$levels
points (which (onraster), raster$x [onraster], col = "blue", pch = 20)
x <- c (sample (1:20, 10), (0 : 5) + 0.45)
raster <- fitraster (x)
raster
plot (x)
abline (h = raster$levels, col = "#00000040")
## unoccupied levels
missing <- setdiff (raster$levels, raster$x)
abline (h = missing, col = "red")
## points acutally on the raster
onraster <- raster$x %in% raster$levels
points (which (onraster), raster$x [onraster], col = "blue", pch = 20)
# }
Run the code above in your browser using DataLab