Re: Update

[John Darrington]

All of Alan's substantial changes have been pushed.

That is to say, we now have the /PRINT=CLUSTERS command, which I think Andy was 
wanting.  There has also been a big improvement to the way the algorithm works,
resulting in better separated clusters.  And some extra subcommands such as 
/CRITERIA=NOINITIAL

We don't yet have a /SAVE option.  In principle, this should be easy.  It will 
be no different from the /SAVE option in linear REGRESSION but as this seems to 
be the subject of bug #44877, I'm reluctant to add that until we have understood
this bug.

J'



On Wed, Dec 09, 2015 at 04:32:17PM -0600, Alan Mead wrote:
     Actually, I think John's recent modification is the best because it
     seems to solve a problem the older implementation had with sometimes
     assigning the clusters in lope-sided ways.
     
     John, is that true (i.e., your recent modification includes my changes,
     right?)?  And do you need anything else? I have supplied some test data
     that I think one of you turned into a test case? And a modification to
     the instructions.  I can re-send any of the bits that were mine.  The
     best way to do this is to upload patches to the bug tracker?
     
     -Alan
     
     
     On 12/9/2015 3:56 PM, Ben Pfaff wrote:
     > It's not in.  I think that we are awaiting an update from Alan.  If
     > not, then I apologize for holding up the process, and I'd appreciate
     > it if Alan could point me to an updated version.
     >
     > On Wed, Dec 9, 2015 at 1:24 PM, Andy King <address@hidden> wrote:
     >> Hi Alan/John,
     >>
     >> Hope both of you are well.  I am just wondering if there is any progress
     >> with this PSPP clustering functionality please?
     >>
     >> Many thanks,
     >>
     >> Andy
     >>
     >>
     >>> Subject: Re: Update
     >>> To: address@hidden
     >>> CC: address@hidden; address@hidden
     >>> From: address@hidden
     >>> Date: Tue, 22 Sep 2015 15:47:41 -0500
     >>> Ok, I might need some help with the test.
     >>>
     >>> -Alan
     >>>
     >>>
     >>> On 9/22/2015 3:12 PM, John Darrington wrote:
     >>>> Hi Alan,
     >>>>
     >>>> My recollection was that we were just waiting for a couple of things,
     >>>> viz:
     >>>> an update to the documentation, and a test for the new subcommand.
     >>>>
     >>>> Once we have this, we can check it in.
     >>>>
     >>>> Regards,
     >>>>
     >>>>
     >>>> John
     >>>>
     >>>> On Tue, Sep 22, 2015 at 02:40:56PM -0500, Alan Mead wrote:
     >>>> Ben or John,
     >>>>
     >>>> Has my patch for the CLUSTER /SAVE option gone into the main release?
     >>>> If not, do we know when it will?
     >>>>
     >>>> At one point, we were waiting for me to fix a bug (the cluster
     >>>> memberships reported for each case needed a translation to match the
     >>>> clusters listed in the main/cluster membership portion of the report).
     >>>> That has been resolved and I emailed a patch. I can submit another
     >>>> patch if necessary.
     >>>>
     >>>> Thanks!
     >>>>
     >>>> -Alan
     >>>>
     >>>>
     >>>> On 9/22/2015 2:09 PM, Andy King wrote:
     >>>>> Hi Alan,
     >>>>>
     >>>>> Thanks for this update, it is greatly appreciated.
     >>>>>
     >>>>> I understand the complexities of adding this to PSPP. Please do keep
     >>>>> me updated as and when things progress.
     >>>>>
     >>>>> Do you think the /PRINT option will be available in the next release
     >>>>> (if it isn't already)?
     >>>>>
     >>>>> Thanks very much,
     >>>>>
     >>>>> Andy
     >>>>>
     >>>>>
     >>>>>
     >>>>> 
------------------------------------------------------------------------
     >>>>> Subject: Re: Update
     >>>>> To: address@hidden
     >>>>> From: address@hidden
     >>>>> Date: Mon, 21 Sep 2015 12:36:50 -0500
     >>>>>
     >>>>> Andy,
     >>>>>
     >>>>> Sorry, I must have missed your email. The short answer is that I
     >>>>> think this feature will become available, but not right away. I'm
     >>>>> fluent in some other programming languages, but not C so coding for
     >>>>> PSPP is difficult. Also, the actual operations involved with the
     >>>>> /SAVE option depend on using internal software features of PSPP
     >>>>> ("readers"). the /PRINT option was a bit simpler.
     >>>>>
     >>>>> I hope this update helps. I appreciate your interest and I wish I
     >>>>> could be more confident in a more urgent solution.
     >>>>>
     >>>>> -Alan
     >>>>>
     >>>>>
     >>>>> On 9/21/2015 12:32 PM, Andy King wrote:
     >>>>>
     >>>>> Hi Alan
     >>>>>
     >>>>> I don't think I've received a response to my below email - any
     >>>>> form of update would be really very helpful and I'd be grateful.
     >>>>>
     >>>>> Many thanks
     >>>>>
     >>>>> Andy
     >>>>>
     >>>>>
     >>>>> 
------------------------------------------------------------------------
     >>>>> From: address@hidden <mailto:address@hidden>
     >>>>> To: address@hidden <mailto:address@hidden>
     >>>>> Subject: RE: address@hidden
     >>>>> <mailto:address@hidden> has sent you a file via WeTransfer
     >>>>> Date: Sat, 12 Sep 2015 21:22:12 +0100
     >>>>>
     >>>>> Hi Alan,
     >>>>>
     >>>>> Hope all is well.
     >>>>>
     >>>>> I am by no means pushing here as I know this is all voluntary but
     >>>>> just wondering if any progress has been made (or news discovered)
     >>>>> on the clustering function looking to be in-built to PSPP?
     >>>>> Specifically the /SAVE command we discussed.
     >>>>>
     >>>>> You may recall we were writing a book....we have now submitted the
     >>>>> final copy of the text for the book chapter we have co-authored.
     >>>>> However, as some of this is online content, I will be able to
     >>>>> factor in this for PSPP if we have any idea when the /SAVE command
     >>>>> will be available? At present it uses SPSS (it would be great to
     >>>>> show off PSPP).
     >>>>>
     >>>>> Cheers!
     >>>>>
     >>>>> A
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>> 
------------------------------------------------------------------------
     >>>>> Subject: Re: FW: address@hidden
     >>>>> <mailto:address@hidden> has sent you a file via WeTransfer
     >>>>> To: address@hidden <mailto:address@hidden>;
     >>>>> address@hidden <mailto:address@hidden>
     >>>>> From: address@hidden <mailto:address@hidden>
     >>>>> Date: Mon, 31 Aug 2015 20:50:22 -0500
     >>>>>
     >>>>> Harry,
     >>>>>
     >>>>> What's new in this version?
     >>>>>
     >>>>> Andy: AFAIK, no additional work on clustering has been done
     >>>>> (unless John's done it).
     >>>>>
     >>>>> -Alan
     >>>>>
     >>>>>
     >>>>> On 8/31/2015 8:04 PM, Andy King wrote:
     >>>>>
     >>>>> Hi Alan
     >>>>>
     >>>>> Is this patch for saving the cluster membership or is it to
     >>>>> correct the csv file handling we can across the other day?
     >>>>>
     >>>>> Thanks
     >>>>>
     >>>>> Andy
     >>>>>
     >>>>>
     >>>>> 
------------------------------------------------------------------------
     >>>>> Date: Mon, 31 Aug 2015 15:44:35 +0000
     >>>>> To: address@hidden <mailto:address@hidden>
     >>>>> From: address@hidden <mailto:address@hidden>
     >>>>> Subject: address@hidden
     >>>>> <mailto:address@hidden> has sent you a file via WeTransfer
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>> address@hidden <mailto:address@hidden>
     >>>>> sent you some files
     >>>>> ?Hi
     >>>>>
     >>>>> In this version there is a little memory leak, but it contains
     >>>>> alan's patch and you can read csv files with it.?
     >>>>>
     >>>>> 
<https://www.wetransfer.com/downloads/daff587c3c2e83198a3cbe713806904c20150831154144/407aa72efc388b1ebf4b31503928dc9920150831154144/c572a6>
     >>>>> Download
     >>>>>
     >>>>> 
<https://www.wetransfer.com/downloads/daff587c3c2e83198a3cbe713806904c20150831154144/407aa72efc388b1ebf4b31503928dc9920150831154144/c572a6>
     >>>>>
     >>>>> 
<https://www.wetransfer.com/downloads/daff587c3c2e83198a3cbe713806904c20150831154144/407aa72efc388b1ebf4b31503928dc9920150831154144/c572a6>
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>> Files (45.6 MB total)
     >>>>> pspp-085+20150827-snapshot-32bits-setup.exe
     >>>>> pspp-085+20150827-snapshot-64bits-setup.exe
     >>>>> Will be deleted on
     >>>>> 7 September, 2015
     >>>>>
     >>>>>
     >>>>> Get more out of WeTransfer, get Plus
     >>>>> <https://www.wetransfer.com/plus?trk=email-recipient_download>
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>> About WeTransfer <https://www.wetransfer.com/about> Contact
     >>>>> <https://www.wetransfer.com/contact> Legal
     >>>>> <https://www.wetransfer.com/legal> _Powered by Amazon Web
     >>>>> Services_
     >>>>> <http://ox-d.wetransfer.com/w/1.0/ai?auid=443977&cs=23423>
     >>>>>
     >>>>>
     >>>>>
     >>>>>
     >>>>> --
     >>>>>
     >>>>> Alan D. Mead, Ph.D.
     >>>>> President, Talent Algorithms Inc.
     >>>>>
     >>>>> science + technology = better workers
     >>>>>
     >>>>> +815.588.3846 (Office)
     >>>>> +267.334.4143 (Mobile)
     >>>>>
     >>>>> http://www.alanmead.org
     >>>>>
     >>>>> Announcing the Journal of Computerized Adaptive Testing (JCAT), a
     >>>>> peer-reviewed electronic journal designed to advance the science and
     >>>>> practice of computerized adaptive testing: http://www.iacat.org/jcat
     >>>>>
     >>>>>
     >>>>> --
     >>>>>
     >>>>> Alan D. Mead, Ph.D.
     >>>>> President, Talent Algorithms Inc.
     >>>>>
     >>>>> science + technology = better workers
     >>>>>
     >>>>> +815.588.3846 (Office)
     >>>>> +267.334.4143 (Mobile)
     >>>>>
     >>>>> http://www.alanmead.org
     >>>>>
     >>>>> Announcing the Journal of Computerized Adaptive Testing (JCAT), a
     >>>>> peer-reviewed electronic journal designed to advance the science and
     >>>>> practice of computerized adaptive testing: http://www.iacat.org/jcat
     >>>> --
     >>>>
     >>>> Alan D. Mead, Ph.D.
     >>>> President, Talent Algorithms Inc.
     >>>>
     >>>> science + technology = better workers
     >>>>
     >>>> +815.588.3846 (Office)
     >>>> +267.334.4143 (Mobile)
     >>>>
     >>>> http://www.alanmead.org
     >>>>
     >>>> Announcing the Journal of Computerized Adaptive Testing (JCAT), a
     >>>> peer-reviewed electronic journal designed to advance the science and
     >>>> practice of computerized adaptive testing: http://www.iacat.org/jcat
     >>>>
     >>>>
     >>> --
     >>>
     >>> Alan D. Mead, Ph.D.
     >>> President, Talent Algorithms Inc.
     >>>
     >>> science + technology = better workers
     >>>
     >>> +815.588.3846 (Office)
     >>> +267.334.4143 (Mobile)
     >>>
     >>> http://www.alanmead.org
     >>>
     >>> Announcing the Journal of Computerized Adaptive Testing (JCAT), a
     >>> peer-reviewed electronic journal designed to advance the science and
     >>> practice of computerized adaptive testing: http://www.iacat.org/jcat
     >>>
     
     -- 
     
     Alan D. Mead, Ph.D.
     President, Talent Algorithms Inc.
     
     science + technology = better workers
     
     +815.588.3846 (Office)
     +267.334.4143 (Mobile)
     
     http://www.alanmead.org
     
     Announcing the Journal of Computerized Adaptive Testing (JCAT), a
     peer-reviewed electronic journal designed to advance the science and
     practice of computerized adaptive testing: http://www.iacat.org/jcat

-- 
Avoid eavesdropping.  Send strong encryted email.
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://sks-keyservers.net or any PGP keyserver for public key.

signature.asc
Description: Digital signature