Usage: |
res = SPKRsurfgls(np, covmod, xmat, nx, dval, alpha, se {, D})
|
Input: |
| np | scalar, degree of the polynomial surface, an integer
in the range 0..6
|
| covmod | integer, type of spatial covariance function:
0 = expcov, 1 = gaucov, 2 = sphercov
|
| xmat | n x 3 matrix, locations (x_i, y_i) [columns 1 & 2] and
observations z_i [column 3]
|
| nx | integer, number of bins used for the table of covariances;
increasing nx adds accuracy, but also increases
the size of the object
|
| dval | scalar, range parameter (for spatial covariance function)
|
| alpha | scalar, proportion of nugget effect (for spatial covariance
function)
|
| se | scalar, standard deviation at distance zero (for spatial
covariance function)
|
| D | optional scalar, dimension of spheres; default D = 2
(only used for spatial covariance function SPKRsphercov)
|
Output: |
| res | list, consisting of components
x, y, z, np, f, alph, l, r, beta, wz, yy, W, l1f,
minx, maxx, covmod and type: |
| res.x | n x 1 vector, same as xmat[,1] |
| res.y | n x 1 vector, same as xmat[,2] |
| res.z | n x 1 vector, same as xmat[,3] |
| res.np | scalar, same as input value np |
| res.f | matrix, internal use only |
| res.alph | matrix, internal use only |
| res.l | matrix, internal use only |
| res.r | matrix, internal use only |
| res.beta | matrix, containing (np + 1)(np + 2) / 2 coefficients |
| res.wz | matrix, internal use only |
| res.yy | matrix, internal use only |
| res.W | matrix, internal use only |
| res.l1f | matrix, internal use only |
| res.minx | 1 x 3 vector, minimum of columns of xmat |
| res.maxx | 1 x 3 vector, maximum of columns of xmat |
| res.covmod | scalar, same as input value covmod |
| res.type | string, "trgls" |