This package implements nonparametric methods for inference on multidimensional spatial (or spatio-temporal) processes, which may be (especially) useful in (automatic) geostatistical modeling and interpolation.
For more information visit the web page of the package.
Nonparametric methods for inference on both spatial trend and variogram functions:
-
np.fitgeo()
(automatically) fits an isotropic nonparametric geostatistical model by estimating the trend and the variogram (using a bias-corrected estimator) iteratively (by callingh.cv()
,locpol()
,np.svariso.corr()
andfitsvar.sb.iso()
at each iteration) -
locpol()
,np.den()
andnp.svar()
use local polynomial kernel smoothing to compute nonparametric estimates of a multidimensional regression function (e.g. a spatial trend), a probability density function or a semivariogram (or their first derivatives), respectively. Estimates of these functions can be constructed for any dimension (depending on the amount of available memory). -
np.svariso.corr()
computes a bias-corrected nonparametric semivariogram estimate using an iterative algorithm similar to that described in Fernandez-Casal and Francisco-Fernandez (2014). This procedure tries to correct the bias due to the direct use of residuals (obtained, in this case, from a nonparametric estimation of the trend function) in semivariogram estimation. -
fitsvar.sb.iso()
fits a ‘nonparametric’ isotropic Shapiro-Botha variogram model by WLS. Currently, only isotropic semivariogram estimation is supported.
Nonparametric residual kriging (sometimes called external drift kriging):
-
np.kriging()
computes residual kriging predictions
(and the corresponding simple kriging standard errors). -
kriging.simple()
computes simple kriging predictions and standard errors. -
Currently, only global (residual) simple kriging is implemented.
Users are encouraged to usegstat::krige()
(orgstat::krige.cv()
) together withas.vgm()
for local kriging.
Among the other functions intended for direct access by the user, the
following (methods for multidimensional linear binning, local polynomial
kernel regression, density or variogram estimation) could be emphasized:
binning()
, bin.den()
, svar.bin()
, h.cv()
and interp()
. There
are functions for plotting data joint with a legend representing a
continuous color scale (based on fields::image.plot()
):
-
splot()
allows to combine a standard R plot with a legend. -
spoints()
,simage()
andspersp()
draw the corresponding high-level plot with a legend strip for the color scale.
There are also some functions which can be used to interact with other
packages. For instance, as.variogram()
(geoR) or as.vgm()
(gstat).
See the Reference for the complete list of functions.
npsp
is available from CRAN, but you can install the development
version from github with:
# install.packages("devtools")
devtools::install_github("rubenfcasal/npsp")
Ruben Fernandez-Casal (Dep. Mathematics, University of A Coruña, Spain). Please send comments, error reports or suggestions to [email protected].
Important suggestions and contributions to some techniques included here were made by Sergio Castillo-Páez (Universidad de las Fuerzas Armadas ESPE, Ecuador) and Tomas Cotos-Yañez (Dep. Statistics, University of Vigo, Spain).
-
Castillo-Páez S., Fernández-Casal R. and García-Soidán P. (2019), A nonparametric bootstrap method for spatial data, 137, Comput. Stat. Data Anal., 1-15, DOI.
-
Fernández-Casal R., Castillo-Páez S. and Francisco-Fernández M. (2018), Nonparametric geostatistical risk mapping, Stoch. Environ. Res. Ris. Assess., 32, 675-684, DOI.
-
Fernández-Casal R., Castillo-Páez S. and García-Soidán P. (2017), Nonparametric estimation of the small-scale variability of heteroscedastic spatial processes, Spa. Sta., 22, 358-370, DOI.
-
Fernandez-Casal R. and Francisco-Fernandez M. (2014) Nonparametric bias-corrected variogram estimation under non-constant trend, Stoch. Environ. Res. Ris. Assess., 28, 1247-1259, DOI.
-
Fernandez-Casal R., Gonzalez-Manteiga W. and Febrero-Bande M. (2003) Flexible Spatio-Temporal Stationary Variogram Models, Statistics and Computing, 13, 127-136, DOI.
-
Rupert D. and Wand M.P. (1994) Multivariate locally weighted least squares regression. The Annals of Statistics, 22, 1346-1370.
-
Shapiro A. and Botha J.D. (1991) Variogram fitting with a general class of conditionally non-negative definite functions. Computational Statistics and Data Analysis, 11, 87-96.
-
Wand M.P. (1994) Fast Computation of Multivariate Kernel Estimators. Journal of Computational and Graphical Statistics, 3, 433-445.
-
Wand M.P. and Jones M.C. (1995) Kernel Smoothing. Chapman and Hall, London.