Description Usage Arguments Details References Examples
Returns the fitting of the regression models at a target site with the best calibration settings.
1 2 3 4 5 6 7 8 |
form |
Formula or list of formulas to try. |
x |
Data frame containing all variables in the formulas. |
x0 |
Data.frame or list containing the input variables of the target |
nk |
Scalar or vector providing a series of neighborhood sizes to try. |
S |
Sampling covariance matrix for the output variable. |
distance |
Vector of distance between the donors and target |
criteria |
Criteria to select the best calibration. One of 'gcv','vp0','avp', 'avpo'. |
lambda |
Scalar that modified the penalty in the GCV criteria |
sigTol |
Lower limit for the model variance during the clalibration |
verbose |
If true a progress bar is printed. |
The GLS model was proposed by Stedinger and Takser (1985):
y= X β + η + ε
where η and ε are two independent terms of errors. The total error is ε = η + δ and includes respectively the error due to sampling, with known covariance matrix S, and a iid term of error η with variance σ^2. Therefore, the covariance matrix of the total error is then given by
Λ = σ^2 I + S = σ^2 G.
For a given model error σ^2, the cholesky decomposition provide the matrix decomposition G^{-1} = UU'. Solving the initial GLS model with known model variance is then equivant to solving the ordinary least-squares problem
U'y = UX + ε^\ast.
A new estimation of the model error can be obtained by
\hatσ^2 = (n-p)^{-1} ∑_{i =1 }^n ε_i^\ast
where n is the number of sites and p the number of parameters. Subsequent iterations are then performed to improved the global fitting of the GLS model until σ^2 has converged.
Stedinger, J. R., & Tasker, G. D. (1985). Regional Hydrologic Analysis: 1. Ordinary, Weighted, and Generalized Least Squares Compared. Water Resources Research, 21(9), 1421–1432. https://doi.org/10.1029/WR021i009p01421
Reis, D. S., Stedinger, J. R., & Martins, E. S. (2005). Bayesian generalized least squares regression with application to log Pearson type 3 regional skew estimation. Water Resources Research, 41(10), W10419. https://doi.org/10.1029/2004WR003445
Kjeldsen, T. R., & Jones, D. A. (2009). An exploratory analysis of error components in hydrological regression modeling. Water Resources Research, 45(2), W02407. https://doi.org/10.1029/2007WR006283
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.