Mailing-List: contact java-dev-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-dev@lucene.apache.org
Message-ID: <1288477591.314361266350968339.JavaMail.jira@brutus.apache.org>
Date: Tue, 16 Feb 2010 20:09:28 +0000 (UTC)
From: "Robert Muir (JIRA)" <jira@apache.org>
To: java-dev@lucene.apache.org
Subject: [jira] Issue Comment Edited: (LUCENE-2091) Add BM25 Scoring to
 Lucene
In-Reply-To: <2076801460.1258983399651.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/LUCENE-2091?page=3Dcom.atlassia=
n.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D128=
34421#action_12834421 ]=20

Robert Muir edited comment on LUCENE-2091 at 2/16/10 8:09 PM:
--------------------------------------------------------------

Joaquin, have you seen this paper: http://doc.rero.ch/lm.php?url=3D1000,43,=
4,20091218142456-GY/Dolamic_Ljiljana_-_When_Stopword_Lists_Make_the_Differe=
nce_20091218.pdf

Its of interest how they modified BM25's idf formula slightly in a way to i=
mprove results when no stopwords list is used. I'm curious what you think a=
bout this as it looks like a potential improvement for people not using sto=
pwords (multilingual situation, etc)

edit here is the quote: for simplicity

{noformat}
Using the original idf formula
idf =3Dlog[(n=E2=88=92dfj +0.5)/(dfj +0.5)], we have noticed
that when the underlying term tj occurs in more than half of
the documents (dfj >n/2), the resulting idf value would be
negative, and the final document score also could be negative.
As a means of estimating idf,we therefore suggest a new variant
defined as idf =3Dlog{1+[(n=E2=88=92dfj +0.5)/(dfj +0.5)]}.
{noformat}


      was (Author: rcmuir):
    Joaquin, have you seen this paper: http://doc.rero.ch/lm.php?url=3D1000=
,43,4,20091218142456-GY/Dolamic_Ljiljana_-_When_Stopword_Lists_Make_the_Dif=
ference_20091218.pdf

Its of interest how they modified BM25's idf formula slightly in a way to i=
mprove results when no stopwords list is used. I'm curious what you think a=
bout this as it looks like a potential improvement for people not using sto=
pwords (multilingual situation, etc)

 =20
> Add BM25 Scoring to Lucene
> --------------------------
>
>                 Key: LUCENE-2091
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2091
>             Project: Lucene - Java
>          Issue Type: New Feature
>          Components: contrib/*
>            Reporter: Yuval Feinstein
>            Priority: Minor
>             Fix For: 3.1
>
>         Attachments: LUCENE-2091.patch, persianlucene.jpg
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> http://nlp.uned.es/~jperezi/Lucene-BM25/ describes an implementation of O=
kapi-BM25 scoring in the Lucene framework,
> as an alternative to the standard Lucene scoring (which is a version of m=
ixed boolean/TFIDF).
> I have refactored this a bit, added unit tests and improved the runtime s=
omewhat.
> I would like to contribute the code to Lucene under contrib.=20

--=20
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-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org