R/outly.R

In dave: Functions for "Data Analysis in Vegetation Ecology"

outly<- function(veg,thresh=0.2,y=0.5){
    dmm<-ncol(veg)
    dc<- as.matrix(as.dist((1-cor(t(veg^0.5)))/2))
    diag(dc)<- 100
    neighb<- apply(dc,1,min)
    neighindex<- apply(dc,1,which.min)
    relevenames<- rownames(veg)
    threshv<- rep(thresh,nrow(veg))
    neighname<- relevenames[neighindex]
    diag(dc)<- 0
    o.pco<- pco(dc)
    # reduce data
    veg.red<- veg[neighb <= thresh,]
    # remove empty species
    rf<- apply(sign(veg.red),2,sum)
    veg.red<- veg.red[,rf > 0]
    o.outlier<- list(threshold=thresh,y=y,rel.names=relevenames,neigh.names=neighname,neigh.dist=neighb,olddim=dim(veg),newdim=dim(veg.red),new.data=veg.red,pco.points=o.pco$points)
}