cv.gsvcm: k-fold cross-validation MSPE for generalized spatially...
In funstatpackages/gsvcm: Generalized Spatially Varying Coefficient Models

Description Usage Arguments Details Value Examples

cv.gsvcm implements k-fold cross-validation MSPE for generalized spartially varying coefficient regression, and returns the mean squared prediction error (MSPE).

1
2
3

cv.gsvcm(y, X, S, V, Tr, d = 2, r = 1, lambda = 10^seq(-6, 6, by =
  0.5), family, off = 0, r.theta = c(2, 8), nfold = 10,
  initial = 123, eps.sigma = 0.01, method = "GSVCM", Cp = TRUE)

`y`	The response of dimension `n` by one, where `n` is the number of observations.
`X`	The design matrix of dimension `n` by `p`, with an intercept. Each row is an observation vector.
`S`	The cooridinates of dimension `n` by two. Each row is the coordinates of an observation.
`V`	The `N` by two matrix of vertices of a triangulation, where `N` is the number of vertices. Each row is the coordinates for a vertex.
`Tr`	The triangulation matrix of dimention `nT` by three, where `nT` is the number of triangles in the triangulation. Each row is the indices of vertices in `V`.
`d`	The degree of piecewise polynomials – default is 2.
`r`	The smoothness parameter – default is 1, and 0 ≤ `r` < `d`.
`lambda`	The vector of the candidates of penalty parameter – default is grid points of 10 to the power of a sequence from -6 to 6 by 0.5.
`family`	The family object, specifying the distribution and link to use.
`off`	offset – default is 0.
`r.theta`	The endpoints of an interval to search for an additional parameter `theta` for negative binomial scenario – default is c(2,8).
`nfold`	The number of folds – default is 10. Although `nfold` can be as large as the sample size (leave-one-out CV), it is not recommended for large datasets. Smallest value allowable for `nfolds` is 3.
`initial`	The seed used for cross-validation sample – default is 123.
`eps.sigma`	Error tolerance for the Pearson estimate of the scale parameter, which is as close as possible to 1, when estimating an additional parameter `theta` for negative binomial scenario – default is 0.01.
`method`	GSVCM or GSVCMQR. GSVCM is based on Algorithm 1 in Subsection 3.1 and GSVCMQR is based on Algorithm 2 in Subsection 3.2 – default is GSVCM.
`Cp`	TRUE or FALSE. There are two modified measures based on the QRGSVCM method for smoothness parameters in the manuscript. TRUE is for Cp measure and FALSE is for GCV measure.