Functional Geostatistics: Univariate and Multivariate Functional Spatial Prediction
Leave-One-Out Cross-Validation for Functional Kriging
Optimal Spatial Design For Functional Data
Classification Function for Functional Data
Leave-One-Out Cross-Validation for Functional cokriging
Functional cokriging
Coordinates of measurement stations Bogota, Colombia
Coordinates of air quality data of Mexico
Coordinates of air quality data of Mexico
Create Covariance Matrices given a series of spatial model parameters
Creates univariate and multivariate CrossSpatFD object to perform cros...
Creates functional ortogonal basis as fd object.
Divide the data in train and test dataset
Generate Variograms for Functional Data from a gfdata object
Creates gfdata objects.
Map plot of a 'KS_pred' object
ggplot of predicted functions
Functional Kriging
map of Bogota, Colombia
map of Mexico
Get the mean of means for each class
Calculate Mean Functions for Each Class
Air quality data of Mexico
Air quality data of Mexico
PM10 of Bogota, Colombia
Print of OptimalSpatialDesign objects
Linear combinations for functional kriging
Spatial random field of scores
Simulation of unconditional or conditional functional spatial process.
Creates univariate and multivariate SpatFD objects.
Summary of COKS_pred objects
Summary of gfdata objects
Summary of KS_pred objects
Summary of SpatFD objects
Coordinates of electrodes from the vowels data set
Performance of functional kriging, cokriging, optimal sampling and simulation for spatial prediction of functional data. The framework of spatial prediction, optimal sampling and simulation are extended from scalar to functional data. 'SpatFD' is based on the Karhunen-Loève expansion that allows to represent the observed functions in terms of its empirical functional principal components. Based on this approach, the functional auto-covariances and cross-covariances required for spatial functional predictions and optimal sampling, are completely determined by the sum of the spatial auto-covariances and cross-covariances of the respective score components. The package provides new classes of data and functions for modeling spatial dependence structure among curves. The spatial prediction of curves at unsampled locations can be carried out using two types of predictors, and both of them report, the respective variances of the prediction error. In addition, there is a function for the determination of spatial locations sampling configuration that ensures minimum variance of spatial functional prediction. There are also two functions for plotting predicted curves at each location and mapping the surface at each time point, respectively. References Bohorquez, M., Giraldo, R., and Mateu, J. (2016) <doi:10.1007/s10260-015-0340-9>, Bohorquez, M., Giraldo, R., and Mateu, J. (2016) <doi:10.1007/s00477-016-1266-y>, Bohorquez M., Giraldo R. and Mateu J. (2021) <doi:10.1002/9781119387916>.