Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: pass (athena.apache.org: domain of pasalic.zaharije@gmail.com
 designates 209.85.216.169 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:date:message-id:subject:from:to:content-type;
        b=dSnAkBGDRo1uKHTWAUDmUpkPxZKCEZLMFuJl9W97wiChaOzwH2poyHF2DO3btPO99T
         Y5DwLDmkzpOhZCliph7JWeNKeYa1Ji6mZJjxzAkpQDh3UeAgbgNRjOxolQPKVwSbBag3
         arGeFTYMXH4BQpHK/2U2BP71FJ3gZWbWU0Oyo=
MIME-Version: 1.0
Date: Fri, 15 Oct 2010 10:30:28 +0200
Message-ID: <AANLkTi=9Y--CrThJWHGMs1FJofwV9_hVnx-8-GBCJHHz@mail.gmail.com>
Subject: Overriding DefaultScore
From: Zaharije Pasalic <pasalic.zaharije@gmail.com>
To: java-user@lucene.apache.org
Content-Type: text/plain; charset=ISO-8859-1

Hi

my original problem is to index large number of documents which
contains 360 integers in rage from 0-90K. Searching it's a little bit
complicated - I need to find most similar documents where query data
is also 360 numbers in range 0-90K. But (there is always 'but') i need
to create score with some predefined weight table. Here is example:

Index contains:

DOC1 : 1, 3, 5
DOC2 : 1, 100
DOC3 : 1, 5

I need to find all documents which are 'like' this:

SEARCH: 1,5,100

And suppose that i'm having table which says: "if value is larger than
10 wight hit as 0.5, else as 1" (in real application this is more
complicated weight table).

So for Query 1,5,100 i will have:

DOC1: SCORE=2    [1,5]
DOC3: SCORE=2    [1,5]
DOC2: SCORE=1.5 [1,100 (100>10- wight 0.5]

Searching is just: if hits occurs on field, increments score by 1*weight(value)

My first step was to create index with one field which contains all
360 values and to remove normals from it.

Now when i'm doing search like:

"F:1 F:5 F:100"

I'm getting results ok but score is not correct. Of course it gives me
score sorted by 'number of hits' (am I right?) but score value is not
calculated by increments of 1 nor i'm using wights at all.

So, my question is - is this even possible with lucene and if can, can
you point me into some directions (i already looked a little bit at
DefaultSimilarity overriding).

Thanks

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