MGWR to be documented
MGWR(Y,XC,XV,S,H, kernels, type = "GD",model='MGWR', minv = 1,
maxknn = 500, NmaxDist = 6000,SE=FALSE, isgcv, TIME, decay,
interceptv=TRUE,W=NULL,betacor=FALSE,remove_local_outlier=FALSE,
outv=0,doMC=FALSE,ncore=1,Wh=NULL,xratiomin=10e-10)A vector
A matrix with covariates with stationnary parameters
A matrix with covariates with spatially varying parameters
A matrix with variables used in kernel
A vector of bandwidths
A vector of kernel types
Type of Genelarized kernel product ('GD' only spatial,'GDC' spatial + a categorical variable,'GDX' spatial + a continuous variable,'GDT' spatial + a time index, and other combination 'GDXXC','GDTX',...)
A mgwrsar model type (see MGWRSAR)
Minimum number of non null weight
If n >NmaxDist how many column with dense weight matrix (max number of neighbours)
Maximum number of observation for computing dense weight matrix
If standard error are computed
leave one out cross validation, default FALSE.
Use rigth truncated kernel for time index kernel
time decay
Intercept spatially varying, default FALSE
A weight matrix for spatial autocorrelation
Do a tuncation of spatial autocorelation if absolute value larger than 1.
Remove local outlier
A treshold for removing local outlier
doParallel parallelization
Number of cores for parallelization
A matrix of weights for local estimation
A treshold parameters for removing obs with not enough positive weigths for local regression
a list of object for MGWRSAR wrapper