TwoStepBootstrap: Estimation and Inference for Two-Step Estimators
In zhenxie23/SpReg: a Toolkit of Spatial Econometrics with Misaligned Data

Description Usage Arguments Value Methods (by generic) See Also Examples

View source: R/TwoStepBootstrap.R

This function implements the esmation and inference for two-step estimators including Krig-and-regress(OLS), Krig-and-regress(GLS), two-step bootstrap.

TwoStepBootstrap(DatR, VarR, DatY, VarY, variogram.model,
                   is.cov.misspecified, is.den.misspecified,
                   plot.start.value = TRUE, cutoff.R, cutoff.res,
                   start.value.method = 2, projected = FALSE)

## S3 method for class 'TwoStepBootstrap'
summary(object, ...)

`DatR`	explanatory variable R, a spatial object, see coordintes()
`VarR`	name of variable R
`DatY`	outcome variable Y, a spatial object, see coordintes()
`VarY`	name of variable Y
`variogram.model`	variogram model type, e.g. "Exp", "Sph", "Gau", "Mat"
`is.cov.misspecified`	logical; if TRUE, the covariance function is misspecified
`is.den.misspecified`	logical; if TRUE, the density function is not Gaussian distribution.
`plot.start.value`	logical, if TRUE, plot the variogram and the fitted variogram curve corresponding to starting values
`cutoff.R`	cutoff for sample variogram of variable R
`cutoff.res`	cutoff for sample variogram of regression residuals
`start.value.method`	fitting method, see fit.variogram()
`projected`	logical; if FALSE, data are assumed to be unprojected, meaning decimal longitude/latitude. For projected data, Euclidian distances are computed, for unprojected great circle distances(km) are computed.
`object`	class `TwoStepBootstrap` objects.

`num.obs`	the number of observations
`vario.par.point.est`	point estimates for variogram parameters(psill, range)
`vario.par.var.mat`	estimated variance-covariance matrix for variogram parameters(psill, range)
`ols.point.est`	point estimates for Krig-and-OLS estimator
`ols.var.mat`	estimated variance-covariance matrix for Krig-and-OLS estimator
`gls.point.est`	point estimates for Krig-and-GLS estimator
`gls.var.mat`	estimated variance-covariance matrix for Krig-and-GLS estimator
`tsbs.point.est`	point estimates for Two-Step Bootstrap estimator
`tsbs.var.mat`	estimated variance-covariance matrix for Two-Step Bootstrap estimator

summary: summary method for class "TwoStepBootstrap".

sp, gstat

library(SpReg)
rivers <- read.csv(system.file("extdata", "rivers.csv", package = "SpReg"))

rivers <- rivers[which(rivers$FOR_NLCD<100),]
train_set <- sample(1:558,277);
test_set <- setdiff(1:558,train_set);
rivers_train <- rivers[train_set, ];
rivers_test <- rivers[test_set, ];

DatR <- rivers_train[,c("FOR_NLCD", "LAT_DD", "LON_DD")];
DatR$X <- log((DatR$FOR_NLCD)/(100-DatR$FOR_NLCD));
sp::coordinates(DatR) <- ~LON_DD+LAT_DD;
sp::proj4string(DatR) =  "+proj=longlat +datum=WGS84";
DatY <- rivers_test[, c("CL","LAT_DD","LON_DD")];
DatY$Y <- log(DatY$CL);
sp::coordinates(DatY) <- ~LON_DD+LAT_DD;
sp::proj4string(DatY) =  "+proj=longlat +datum=WGS84";

TwoStep_Results <- TwoStepBootstrap(DatR, "X", DatY, "Y", "Exp", FALSE, FALSE,
                                    cutoff.R = 295, cutoff.res = 40);