X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=heikki%2FREADME-HEIKKI;h=0b547a1f93c0beaaaeafb416294f8c12a55bccb2;hb=refs%2Fheads%2Franking-h;hp=05718861e702ee4607ce8fd91d4d41aec2ea3c20;hpb=533ff0b399b62625ddd68e1a59cb34ef70795bf9;p=pazpar2-moved-to-github.git

diff --git a/heikki/README-HEIKKI b/heikki/README-HEIKKI
index 0571886..0b547a1 100644
--- a/heikki/README-HEIKKI
+++ b/heikki/README-HEIKKI
@@ -62,3 +62,28 @@ Add this to the target defs
 After this, it should be possible to get records from different databases, some
 with many records, some with a few. This is a good testing ground for merging
 rankings! Test first with a round-robin, and plot the scores.
+
+Thu 28-Nov
+Ok, I can now merge a number of SOLR databases (harvest jobs), and plot their rankings
+as solr gives them, in the order of different merge strategies
+Next: Add the normalizing merge strategy. Then plot different strategies against different queries
+Write a conclusion, and consider this plotting job done
+
+
+Fri 13-Dec-2013
+Adam is adding a float type to pazpar2. I have made a prrof of concept of the normalizing
+by curve fitting. I think it is time to close this branch, and start (re)implementing
+things in the main branch. Keep the old branch around for reference!
+
+Need new config options:
+ - sort: native, native + position
+ - or per target: native score / fake score from position / use tf/idf
+ - per target: weight for combining rankings (cluster merge), so we can trust one
+   target more than others
+ - per target: boost rankings
+
+Start coding:
+ - in relevance-prepare-read, go through records, collect scores in arrays (per target),
+ - fit the curve, normalize the scores.
+ - cluster scoring
+