[ https://issues.apache.org/jira/browse/LUCENE-8311?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16483764#comment-16483764 ]
Robert Muir commented on LUCENE-8311: ------------------------------------- I wonder if its difficult to test with another similarity such as a DFR model? I'm only asking because I'm a little concerned that the bogus way we compute "phrase IDF" for BM25Similarity & ClassicSimilarity is getting in your way. All the other models use a more sane approach (scores like a disjunction internally). BM25 carried along the brain damage of ClassicSimilarity just because it was trying to minimize differences, but not for any particular good reason. > Leverage impacts for phrase queries > ----------------------------------- > > Key: LUCENE-8311 > URL: https://issues.apache.org/jira/browse/LUCENE-8311 > Project: Lucene - Core > Issue Type: Improvement > Reporter: Adrien Grand > Priority: Minor > Attachments: LUCENE-8311.patch > > > Now that we expose raw impacts, we could leverage them for phrase queries. > For instance for exact phrases, we could take the minimum term frequency for > each unique norm value in order to get upper bounds of the score for the > phrase. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org