Normalization is done on a field by field basis, as is most scoring.
It doesn't factor all fields in, b/c someone might not be querying all
fields. The field it does use is based on the query.
On Aug 18, 2008, at 10:44 PM, blazingwolf7 wrote:
Hi,
I am currently working on the calculation of score part in Lucene.
And I
encounter a part that I do not understand.
return raw * Similarity.decodeNorm(norms[doc]); // normalize for
field
As can be seen from the code above, the Similarity method
decodeNorm() will
be called to decode the byte formatted value to change it back into
float
value. This value actually represent the normalization value for the
fields;
title, url, content, host and anchor.
I would like to know how it actually select the field to be included
in the
calculation. When I print out the value, I noticed only one of the
field
will be selected. Can anyone advise me on this? How the field is
selected,
and why not all the field is used in the calculation?
Thanks
--
View this message in context:
http://www.nabble.com/Score-Boosting-tp19043489p19043489.html
Sent from the Lucene - Java Users mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
