gridweights
Compute Quadrature Weights Based on Grid Counts
Computes quadrature weights for a given set of points, using the ``counting weights'' for a grid of rectangular tiles.
Usage
gridweights(X, ntile, …, window=NULL, verbose=FALSE, npix=NULL, areas=NULL)Arguments
- X
Data defining a point pattern.
- ntile
Number of tiles in each row and column of the rectangular grid. An integer vector of length 1 or 2.
- …
Ignored.
- window
Default window for the point pattern
- verbose
Logical flag. If
TRUE, information will be printed about the computation of the grid weights.- npix
Dimensions of pixel grid to use when computing a digital approximation to the tile areas.
- areas
Vector of areas of the tiles, if they are already known.
Details
This function computes a set of quadrature weights
for a given pattern of points
(typically comprising both ``data'' and `dummy'' points).
See quad.object for an explanation of quadrature
weights and quadrature schemes.
The weights are computed by the ``counting weights'' rule
based on a regular grid of rectangular tiles.
First X and (optionally) window are converted into a
point pattern object. Then the bounding rectangle of the window of
the point pattern is
divided into a regular ntile[1] * ntile[2] grid of rectangular tiles.
The weight attached to a point of X is the area of the tile
in which it lies, divided by the number of points of X lying in
that tile.
For non-rectangular windows the tile areas are currently calculated
by approximating the window as a binary mask. The accuracy of this
approximation is controlled by npix, which becomes
the argument dimyx of as.mask.
Value
Vector of nonnegative weights for each point in X.
See Also
Examples
# NOT RUN {
Q <- quadscheme(runifpoispp(10))
X <- as.ppp(Q) # data and dummy points together
w <- gridweights(X, 10)
w <- gridweights(X, c(10, 10))
# }