Re: Cluster membership in SPSS

[John Darrington]

On Mon, Jul 27, 2015 at 12:11:19PM -0500, Alan Mead wrote:
     
     In the attached files, the .patch is the patch for 
     pspp-0.8.5/src/language/stats/quick-cluster.c (let me know if it's done
     incorrectly)

It looks good to me.  Just some minor quibbles about style:


+      // get cluster membership (does not reflect re-ordered centers)
+      clust = kmeans_get_nearest_group (kmeans, c, qc); 
+      // make cluster membership reflect re-ordered centers

In pspp we don't use // style comments, because older compilers can't
handle them.

+      tab_text_format (t, 0, i+1, TAB_CENTER, "%d", (i + 1));
+      tab_text_format (t, 1, i+1, TAB_CENTER, "%d", (clust + 1));

We like to put spaces around operators.  So write "i + 1" instead of "i+1"



I think this patch could be checked in if we also do the following:


1.  Add the necessary changes to doc/statistics.texi documenting the
   /PRINT subcommand.

2.  Add a test to tests/language/stats/quick-cluster.at which exercises
   the /PRINT subcommand.

3. We should omit the syntax parsing code for the /SAVE command, because
   it doesn't do anything yet.


Otherwise, great work!


J'


     

-- 
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