Gradient-based Stochastic Local Interaction (SLI) Model
SLI_GRADIENT_MLE_2020 IS A SOFTWARE BUNDLE THAT IMPLEMENTS THE
GRADIENT-BASED STOCHASTIC LOCAL INTERACTION SPATIAL MODEL in D dimensions.
EXAMPLES: Examples for SLI parameter estimation, interpolation, and
mapping are given in the script Main_examples.m
The three main functions for SLI estimation, interpolation and mapping
are those having names that begin with "Main":
1. Main_sli_mle: It should be run before the other scripts.
It estimates the SLI parameters for a specified dataset.
2. Main_loocv_sli: This function can be run after Main_sli_mle.m if it
is desired to obtain Leave-One-Out Cross validation estimates.
3. Main_sli_map: This function is executed in order to interpolate a
give dataset and generated predictions at the nodes of a map grid. This
works only for two-dimensional data z(x,y)
4. INTERPOLATION AT SET OF P POINTS
Note that if the target is to predict the unknown values at a number of
sites (not necessarily on a map) one can simply execute the command:
[ Zphat, svar ] = Predict_SLI(dd, locs, zs, plocs, kernel, K, modelsli, Param_sli );
where:
dd: Space dimension
locs: Data locations (array N x dd)
zs: Data values (array N x 1)
plocs: Prediction Locations (array P x dd)
kernel: String specifying kernel function (help kernel_s for
different types)
K: Neighbor order (K=3 should be adequate)
modelsli: SLI model structure (you need to first run Main_sli_mle.m
for the same dataset)
Param_sli: SLI parameter vector (you need to first run Main_sli_mle.m
for the same dataset)
Copyright (C) 2020 Dionisis Hristopulos, Andreas Pavlides, Vasiliki Agou
email: dchristopoulos@tuc.gr
This bundle is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation (version 2 of the License or later version).
This program is distributed WITHOUT ANY WARRANTY.
See the GNU General Public License for more details.
If you use this software, please cite the publication:
Hristopulos, D.T., Pavlides, A., Agou, V.D. et al. Stochastic Local Interaction Model: An Alternative to Kriging for Massive Datasets. Math Geosci 53, 1907–1949 (2021). https://doi.org/10.1007/s11004-021-09957-7