R/weight.time.R

In AEM: Tools to construct Asymmetric eigenvector maps (AEM) spatial variables

weight.time <- function(dates, distmat=NULL,alpha=2, beta=NULL, max.d=NULL, unit.angle="degrees", rot.angle=0, rm.same.y=TRUE, plot.connexions=TRUE){
#
# dates : vector of dates.
# alpha : exponent of the first weighting function
# beta  : exponent of the second weighting function
# max.d : maximum distance for weighting. Default value (not recommended in most 
#         cases): the maximum distance among dates in the 'dates' vector. 
#         A more meaningful solution in many applications is to compute a  
#         Moran's I correlogram (for univariate data) or a Mantel correlogram 
#         (for multivariate data), and use the distance where the correlation 
#         becomes 0 as the value for max.d.
#
### F. Guillaume Blanchet - August 2013
##########################################################################################
	n  <- length(dates)
	nb <- cell2nb(1,n)
	xy <- cbind(1:n, rep(1, n), dates)
	w  <- weight.edges(nb.object=nb, coords=xy, distmat=distmat, alpha=alpha, beta=beta, max.d=max.d, unit.angle=unit.angle, rot.angle=rot.angle, rm.same.y=rm.same.y, plot.connexions=plot.connexions)	
	w
}