GRIDN Function

Standard Library function that grids n dimensional data.

Usage

result = GRIDN(d, i)

Input Parameters

d—An (m,n+1) array of m datapoints in n independent variables and one dependent variable; d(*,n) is the dependent variable.

i—A vector of n integers specifying the dimensions of the grid.

Returned Value

result—An n dimensional array of values of the dependent variable on a regular grid over the independent variables.

Keywords

b—A 2 × n array fixing the boundary of the grid. b(0,*) is the minimum corner and b(1,*) is the maximum corner. The default extent of the grid is the same as that of the data.

c—(output) A list of n vectors defining the grid coordinates.

f—The name of a user-supplied procedure describing voids in the independent variable space (datapoints and gridpoints within these regions are ignored in computation). Input to f is a (p,n) array of p points in the independent variable space. f outputs two items where the first item is a vector of indices indicating which of the p input points are within bounds, and where the second item is a scalar that will appear as a place holder for the dependent variable at out-of-bounds gridpoints.

r—Scalar specifying order of the weighting function. The dependent variable at a grid point is computed as a weighted average of the variable over all neighborhood datapoints. The weighting function is 1/e^r where e is the Euclidean distance between the grid point and the datapoint. Choosing an even integer value for r allows PV‑WAVE to use integer exponentiation, which is substantially faster for larger data sets than floating-point exponentiation. The default value is 2.

s—The name of a user-supplied function that computes distance. The inputs are the (p,n) array u and the (q,n) array v defining two sets of points in the independent variable space. The output is the (p,q) array w where w(j,k) is the square of the distance between points u(j,*) and v(k,*). The default measure of distance is (cartesian) Euclidean.

t—A scalar between 0 and 1 specifying neighborhood size. t=1 gives a maximal neighborhood which includes all datapoints, while lower t values yield smaller neighborhoods. For dense data sets, a value of t = 0 will improve performance by an order of magnitude or more. t defaults to 1

Examples

See wave/lib/user/examples/gridnex1.pro