[ https://issues.apache.org/jira/browse/LUCENE-1896?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12752271#action_12752271 ]
Mark Miller edited comment on LUCENE-1896 at 9/7/09 2:55 PM: ------------------------------------------------------------- Hmm .. so perhaps the idea that the query scores are more comparable comes from: if we take it out, that means the score will be the cosine times the euclidean distance - the distance will be rather large for a small query vector and a large doc vector (in magnitude). So a query that matched many large docs would could scale a lot harder than a query that hit small docs? I guess that makes sense - scores not being very comparable anyway though, it hardly seems worth the extra compute cost ... was (Author: markrmil...@gmail.com): Hmm .. so perhaps the idea that they query scores are more comparable comes from: if we take it out, that means the score will be the cosine X the euclidean distance - the distance will be rather large for a small query vector and a large doc vector (in magnitude). So a query that matched many large docs would could scale a lot harder than a query that hit small docs? I guess that makes sense - scores not being very comparable anyway though, it hardly seems worth the extra compute cost ... > Modify confusing javadoc for queryNorm > -------------------------------------- > > Key: LUCENE-1896 > URL: https://issues.apache.org/jira/browse/LUCENE-1896 > Project: Lucene - Java > Issue Type: Improvement > Components: Javadocs > Reporter: Jiri Kuhn > Priority: Minor > Fix For: 2.9 > > > See http://markmail.org/message/arai6silfiktwcer > The javadoc confuses me as well. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online. --------------------------------------------------------------------- To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org For additional commands, e-mail: java-dev-h...@lucene.apache.org