[R] Plot fail with dbscan (package fpc) and dataset with only noise

Forconi forconi at wavecomm.it
Mon Apr 22 10:36:46 CEST 2013

I'm using the clustering package 'fpc' to make some test with the dbscan algorithm.

The commands I'm using are rather simple:
   csvdata <- read.table("C:/file.csv", header = TRUE)
   dataset <- csvdata[,c('D1', 'D2', 'D3', 'D4', 'D5', 'D6', 'D7')]
   ds <- dbscan(dataset, 0.15, MinPts = 10)
   plot(ds, dataset)

The problem I have is that when the dbscan algorithm doesn't recognize any cluster (but only noise), the plot command fail saying:
Error in axis(side = side, at = at, labels = labels, ...) : 
  invalid value specified for graphical parameter "pch"

And if I try to set a value for pch with the command plot(ds, dataset, pch = 5), it fails with the message:
Error in localWindow(xlim, ylim, log, asp, ...) : 
  formal argument "pch" matched by multiple actual arguments

Is there any method to plot the results when the dbscan algorithm does not recognize any cluster but only noise?

Thanks for any help.


More information about the R-help mailing list