bugGNU Octave - Bugs: bug #50365, [octave forge] (statistics) Wrong...

 
 

bug #50365: [octave forge] (statistics) Wrong distance results from kmeans () and wrong centers

Submitter:  Georg Wiora <gwiora>
Submitted:  Tue 21 Feb 2017 09:06:57 AM UTC
   
 
Category:  Octave Package Severity:  4 - Important
Priority:  7 - High Item Group:  Incorrect Result
Status:  Fixed Assigned to:  juanpi
Originator Name:  gwiora Open/Closed:  * Closed
Release:  * 5.1.0 Operating System:  * Any
Fixed Release:  None Planned Release:  None
* Mandatory Fields

Add a New Comment Rich Markup
   

Thu 11 Apr 2019 12:28:35 PM UTC, comment #2: 

This is fixed in development version.
commit fc75e8b3b37b

Please verify

I am marking this as done.

Juan Pablo Carbajal <juanpi>
Group Member
Thu 11 Apr 2019 12:23:48 PM UTC, comment #1: 

This is still the case for v1.4.1 the problem arises when replicas are used.
The distance matrix and the classes are not updated to the best replica.

Working on a fix

Juan Pablo Carbajal <juanpi>
Group Member
Tue 21 Feb 2017 09:06:57 AM UTC, original submission:  

The kmeans() function from statisics package V1.3.0. does return wrong center positions and wrong center distances.

I am using a 3-d data set and found at least one bug in code review of kmeans.m but fixing it did non give the correct results.

The value returned for "dist" result is probably the sum of squares of coordinate differences. But should be the euclidian distance at least when 'distance'='sqEuclidean' is set. Octave and matlab documentation both state that the result should be euclidian distance.

When comparing cluster centers returned they do not match the points associated with the index results. Something seems to be serverly screwed up here. I will come back here when i can provide a test function with synthetic data.

Georg Wiora <gwiora>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by juanpi (Posted a comment)
  • -email is unavailable- added by jwe (Updated the item)
  • -email is unavailable- added by gwiora (Submitted the item)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only group members can vote.

     

    Follow 8 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2019-04-11 juanpi StatusIn Progress Fixed
        Open/ClosedOpen Closed
    2019-04-11 juanpi Severity3 - Normal 4 - Important
        Priority5 - Normal 7 - High
        StatusNone In Progress
        Assigned toNone juanpi
        Release4.2.0 5.1.0
    2017-08-12 jwe SummaryWrong distance results from kmeans () and wrong centers [octave forge] (statistics) Wrong distance results from kmeans () and wrong centers

    Back to the top

    Powered by Savane 3.13-3230.
    Corresponding source code