[R] Specifying medoids in PAM?

[Martin Maechler]

>>>>> "David" == David Finlayson <david.p.finlayson at gmail.com>
>>>>>     on Wed, 8 Jun 2005 09:24:54 -0700 writes:

    David> Sorry, I wasn't trying to submit a bug report just yet. 

the posting guide asks you to provide reproducible examples, in
any case, not just for bug reports ...
{and strictly speaking, you still haven't provided one, since
 it's a bit painful to read in your table below -- because of the
 extra row names ... but here I'm nit picking a bit }

    David> I wanted to see if I was using the command correctly. 

Yes, you were.

    David> Here is a complete example (the row names have numbers in them):

    >> library(cluster)
    >> stats.table
    David> slope.for slope.ter prof.len break.elv
    David> 1 ALLYN W       0.09      0.05     63.3      1.46
    David> 2 ARCADIA       0.12      0.09     40.8      0.43
    David> 3 BURFOOT       0.12      0.03     58.9      0.78
    David> 4 CAMA BE       0.15      0.05     50.4      0.88
    David> 5 DASH PO       0.09      0.01    290.5      3.14
    David> 6 EAGLE C       0.09      0.01    226.4      1.49
    ..................

    >> pam(stats.table, metric="euclidean", stand=TRUE, medoids=c(1,3,20,2,5), k=5)

    David> This command crashes RGUI.exe and windows sends an error report to
    David> Microsoft. It also crashes if I first subtract the NA rows from
    David> stats.table.

I can confirm to get segmentation faults using this example data
with k=5 ,  so effectively, it seems you've uncovered a bug in pam().
I will investigate and patch eventually.

Thank you for the report.
Martin Maechler, 
ETH Zurich

    David> Thanks,

    David> David 


    David> On 6/8/05, Martin Maechler <maechler at stat.math.ethz.ch> wrote:
    >> >>>>> "David" == David Finlayson <david.p.finlayson at gmail.com>
    >> >>>>>     on Tue, 7 Jun 2005 12:11:25 -0700 writes:
    >> 
    David> I am using the PAM algorithm in the CLUSTER library.
    David> When I allow PAM to seed the medoids using the default __build__
    David> algorithm things work
    David> well:
    >> 
    >> >> pam(stats.table, metric="euclidean", stand=TRUE, k=5)
    >> 
    David> But I have some clusters from a Hierarchical analysis that I would
    David> like to use as seeds for the PAM algorithm. I can't figure what the
    David> mediod argument wants. When I put in the five integer indicies for the
    David> observations in stats.table that I would like to use as seeds (the row
    David> numbers), I segfault R.
    >> 
    >> >> pam(stats.table, metric="euclidean", stand=TRUE, medoids=c(1,3,20,2,5), k=5)
    >> 
    David> *** R Crashes ***
    >> 
    >> this is not very helpful.
    >> 
    >> Can you please   READ the posting guide and then
    >> do as the guide says :
    >> 
    >> -  post a *reproducible* example
    >> -  tell more about what happens when  "R Crashes"
    >> 
    David> Here is my version info:
    >> >> version
    David> _
    David> platform i386-pc-mingw32
    David> arch     i386
    David> os       mingw32
    David> system   i386, mingw32
    David> status
    David> major    2
    David> minor    0.1
    David> year     2004
    David> month    11
    David> day      15
    David> language R
    >> 
    David> Any guidance would be appreciated.
    >> 
    David> David
    >> 
    David> --
    David> David Finlayson
    David> Marine Geology & Geophysics
    David> School of Oceanography
    David> Box 357940
    David> University of Washington
    David> Seattle, WA  98195-7940
    David> USA
    >> 
    David> Office: Marine Sciences Building, Room 112
    David> Phone: (206) 616-9407
    David> Web: http://students.washington.edu/dfinlays
    >> 
    David> ______________________________________________
    David> R-help at stat.math.ethz.ch mailing list
    David> https://stat.ethz.ch/mailman/listinfo/r-help
    David> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
    >> 


    David> -- 
    David> David Finlayson
    David> Marine Geology & Geophysics
    David> School of Oceanography
    David> Box 357940
    David> University of Washington
    David> Seattle, WA  98195-7940
    David> USA

    David> Office: Marine Sciences Building, Room 112
    David> Phone: (206) 616-9407
    David> Web: http://students.washington.edu/dfinlays