[Octave-bug-tracker] [bug #49411] fcm problem with clustering big data sets

[Tony Trew]

URL:
  <http://savannah.gnu.org/bugs/?49411>

                 Summary: fcm problem with clustering big data sets
                 Project: GNU Octave
            Submitted by: tonytrew
            Submitted on: Sat 22 Oct 2016 07:15:50 AM GMT
                Category: Octave Forge Package
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Crash
                  Status: None
             Assigned to: None
         Originator Name: Tony Trew
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: 4.0.2
        Operating System: Microsoft Windows

    _______________________________________________________

Details:

Using fcm in the fuzzy-logic-toolkit-0.4.5 runs into problems with bigger sets
of data points. This is the function call:
 

[cluster_centers, soft_partition, obj_fcn_history]=fcm(Xpc,k,[2,100,1e-4]);


Xpc (size 22613 X 14) is a set of election results in percentages with each
point summing to 1.00 

The script terminates at the first iteration with the message:

Iteration count = 1,  Objective fcn = NaN


If the number of points is reduced to 2700  it completes and returns
[cluster_centers, soft_partition, obj_fcn_history], as well as values for:
partition_coeff (soft_partition));
partition_entropy (soft_partition, 2));
xie_beni_index (Xpc1, cluster_centers, soft_partition));

But any more points than 2700 and the same problem occurs, even if the number
of clusters is reduced to 2 and the dimension of the data points decreased.

I am running this on an HP laptop, with 8.00 Gb RAM, 64-bit operating system ,
x64 based processor












    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?49411>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/