gwglmnet.nen.cv.f: Cross-validation for selection of tuning parameter in a...

In wrbrooks/gwselect: Variable selection in Geographically Weighted Regression models

Description

Cross-validation for selection of tuning parameter in a GW-GLM model using Nearest Effective Neighbors for bandwidth selection.

Usage

gwglmnet.nen.cv.f(formula, data, bw, coords, gweight, verbose, adapt, longlat, s = NULL, beta1, beta2, family, weights = NULL, D = NULL, tolerance = .Machine$double.eps^0.25, type = "pearson", parallel = FALSE, ...)

Arguments

formula
data
bw
coords
gweight
verbose
adapt
longlat
s
beta1
beta2
family
weights
D
tolerance
type
parallel
...

Author(s)

Wesley Brooks

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (formula, data, bw, coords, gweight, verbose, adapt, 
    longlat, s = NULL, beta1, beta2, family, weights = NULL, 
    D = NULL, tolerance = .Machine$double.eps^0.25, type = "pearson", 
    parallel = FALSE, ...) 
{
    cat(paste("Beginning with target SSR: ", bw, ", tolerance: ", 
        tolerance, "\n", sep = ""))
    gwglmnet.model = gwglmnet.nen(formula = formula, data = data, 
        coords = coords, gweight = gweight, bw = bw, verbose = verbose, 
        longlat = longlat, adapt = adapt, s = s, family = family, 
        weights = weights, D = D, tol = tolerance, beta1 = beta1, 
        beta2 = beta2, type, parallel = parallel)
    print(gwglmnet.model[["model"]][["cv.error"]])
    print(names(gwglmnet.model))
    print(gwglmnet.model[["model"]])
    cv.error = sum(sapply(gwglmnet.model[["model"]], function(x) min(x[["cv.error"]])))
    cat(paste("Bandwidth: ", bw, ". CV error: ", cv.error, "\n", 
        sep = ""))
    return(cv.error)
  }

wrbrooks/gwselect documentation built on May 4, 2019, 11:59 a.m.