Dear all, I try to use qtclust() (package flexclust) to do QT (quality threshold) clustering. But I don't know how to set "Pearson correlation " as similarity measure. I would like to know how to set "Minimum correlation" ( for example 0.9) as well. Anyone has any idea? Or is there any other packages for QT cluster can set those parameters? Many thanks in advance. The following is the R code I wrote, library(flexclust) mycont=list(iter.max=400,ntry=nrow(ClusterData),min.size=5) as(mycont,"flexclustControl") c11<-qtclust(ClusterData,radius=1,control=mycont) Best regards, Guiyuan [[alternative HTML version deleted]]

Hi again, I think I know how to set those parameters by constructing "kccaFamily" object now. mycont=list(iter.max=1000,ntry=nrow(ClusterData),min.size=10) as(mycont,"flexclustControl") mykccaFamily<-kccaFamily(dist=distCor,similarity=TRUE) #minimum correlation 0.9 c11<-qtclust(ClusterData,family=mykccaFamily,radius=0.1,control=mycont ) Best regards, Guiyuan On Tue, Jul 29, 2008 at 3:04 PM, Guiyuan Lei <guiyuanlei@gmail.com> wrote: > Dear all, > > I try to use qtclust() (package flexclust) to do QT (quality threshold) > clustering. But I don't know how to set "Pearson correlation " as similarity > measure. I would like to know how to set "Minimum correlation" ( for > example 0.9) as well. Anyone has any idea? Or is there any other packages > for QT cluster can set those parameters? Many thanks in advance. > > The following is the R code I wrote, > > library(flexclust) > mycont=list(iter.max=400,ntry=nrow(ClusterData),min.size=5) > as(mycont,"flexclustControl") > c11<-qtclust(ClusterData,radius=1,control=mycont) > > > Best regards, > Guiyuan > [[alternative HTML version deleted]]