lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Muir <rcm...@gmail.com>
Subject Re: problems writing a custom Similarity class
Date Sat, 01 Oct 2011 01:44:46 GMT
On Sun, Sep 25, 2011 at 1:35 PM, Jason Toy <jasontoy@gmail.com> wrote:
> Scoring seems to still be using an idf score that is not 1 and returning
> results sorted by rareness of a phrase instead of frequency of the word.

also keep in mind overriding idf(int, int) might not give you what you
want for phrase queries (since you said phrase in your email).

by default the idf of a phrase is computed by summing the IDF across
the terms...
so with your idf implementation the idf for a phrase would be equal to
the number of terms in that particular phrase

in lucene 3.4 to change how this is done you need to also override
idfExplain(Collection<Term> terms, Searcher searcher)

-- 
lucidimagination.com

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


Mime
View raw message