Hi Leo,

This is probably not something to be concerned about, since some of
the cluster stopping mechanisms incorporate a bit of randomization,
and so there can be some variation in the results depending on how
that is done. I would guess that the differences aren't a lot though
(usually between 3 or 4 clusters, for example). So, I would not be too
worried about these tests failing. I should also make sure to make
this a bit more robust (both checking for 64 bit systems and then
allowing for a bit of variation in the cluster stopping results.)  The
same issue exists for SVD, btw, where results can vary a bit due to
some randomizations done during SVD - in that case the test cases have
been adapted to be a bit "fuzzier" so usually those run ok.

I hope this helps, let us know as questions arise!

Good luck,
Ted

On Thu, Sep 10, 2009 at 12:50 PM, Leo Alekseyev <[email protected]> wrote:
> Ted, thanks for your response.  All tests now run, however a number of
> tests in the /Testing/clusterstop/clusterstopping directory still
> fail, e.g.:
>
>
> # csh ./testA1.sh
> Test A1 - Testing clusterstopping.pl with default settings.
> Running clusterstopping.pl --prefix testA1 testA1.vectors
> STATUS : ERROR Test Results don't Match....
> When Tested Against testA1.reqd -
> 1c1
> < 8
> ---
>> 3
> STATUS : OK File testA1.cr.dat created.
> STATUS : OK File testA1.pk3 created.
> STATUS : OK File testA1.pk3.dat created.
>
> -----------
>
> # csh ./testA2.sh
> Test A2 - Testing clusterstopping.pl when using all measures.
> Running clusterstopping.pl --prefix testA2 --measure all --clmethod
> rbr --crfun e1 --precision 6 --seed 3 testA2.vectors
> STATUS : ERROR Test Results don't Match....
> When Tested Against testA2.reqd -
> 1c1
> < 10
> ---
>> 9
> STATUS : OK File testA2.cr.dat created.
> STATUS : OK File testA2.pk1 created.
> .................
>
> Is this something to be concerned about?..
> --Leo
>
>
> On Thu, Sep 10, 2009 at 5:06 AM, Ted Pedersen <[email protected]> wrote:
>> If you have a 64 bit machine, you'll want to make sure you are using
>> the 64 bit version of Cluto.
>>
>> http://glaros.dtc.umn.edu/gkhome/fetch/sw/cluto/cluto-2.1.2a.tar.gz
>>
>> The install.sh script that comes with SenseClusters only uses version
>> 2.1.1, which doesn't handle 64 bit Linux, so this is something we will
>> change. However, for now I think if you download the 64 bit version of
>> Cluto that should resolve this problem.
>>
>> More on the different versions of Cluto can be found here:
>>
>> http://glaros.dtc.umn.edu/gkhome/cluto/cluto/download
>>
>> Please let us know if additional questions arise!
>>
>> Good luck,
>> Ted
>>
>> On Wed, Sep 9, 2009 at 9:20 PM, Leo Alekseyev <[email protected]> wrote:
>>> I am trying to install Senseclusters as root on a linux machine
>>> running Ubuntu 9.04 64 bit; everything appears to install fine but the
>>> test suite fails on clusterstopping.pl -- apparently there are
>>> problems with the vcluster executable
>>>
>>> now /usr/local/bin is in PATH and when i try to just type "vcluster"
>>> at the command line it says
>>> -su: /usr/local/bin/vcluster: No such file or directory
>>>
>>> The test script output is below.  I'd appreciate any help with this!
>>> --Leo
>>>
>>> r...@ubuntu:/home/leo/downloads/Text-SenseClusters-1.01/Testing/clusterstop/clusterstopping#
>>> csh ./testA1.sh
>>> Test A1 - Testing clusterstopping.pl with default settings.
>>> Running clusterstopping.pl --prefix testA1 testA1.vectors
>>> sh: vcluster: not found
>>> Error while running vcluster --clmethod rb --crfun i2 --sim cos
>>> --rowmodel none --colmodel none --nooutput  testA1.vectors 1
>>> STATUS : ERROR Test Results don't Match....
>>> When Tested Against testA1.reqd -
>>> 0a1
>>>> 3
>>> STATUS : OK File testA1.cr.dat created.
>>> STATUS : OK File testA1.pk3 created.
>>> STATUS : ERROR File testA1.pk3.dat NOT created.
>>>
>>> ------------------------------------------------------------------------------
>>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
>>> trial. Simplify your report design, integration and deployment - and focus 
>>> on
>>> what you do best, core application coding. Discover what's new with
>>> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
>>> _______________________________________________
>>> senseclusters-users mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/senseclusters-users
>>>
>>
>>
>>
>> --
>> Ted Pedersen
>> http://www.d.umn.edu/~tpederse
>>
>> ------------------------------------------------------------------------------
>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
>> trial. Simplify your report design, integration and deployment - and focus on
>> what you do best, core application coding. Discover what's new with
>> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
>> _______________________________________________
>> senseclusters-users mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/senseclusters-users
>>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> senseclusters-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/senseclusters-users
>



-- 
Ted Pedersen
http://www.d.umn.edu/~tpederse

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
senseclusters-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/senseclusters-users

Reply via email to