lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Hany Azzam <>
Subject Re: IndexSearcher with two Indexes
Date Fri, 27 Jan 2012 21:53:51 GMT
Hi Robert,

Thanks for the reply. I am trying to do something different. If I use a mutireader then the
searching/scoring will take place over the two indexes at the same time. However, in my case
the subcomponents of the retrieval model are calculated over separate evidence spaces. For
example, the retrieval model calculates something like that: 

score := P(query_term | documents) * P(query_term | relevant_documents)

The P(query_term | documents) can be estimated using the index over the whole collection of
documents. The P(query_term | relevant_documents) can be estimated using the index over the
relevant documents only (which are known prior to the execution of the query).

The question is can I do such a calculation which uses to separate indexes in one scoring

Of course one option is to use the MultiSimilarity Class and combine the score somehow. However,
the retrieval function is more complex than that and a simple combination using product or
summation won't be feasible.

Any ideas on how to resolve this problem (if possible :))?

Thanks again,

On 27 Jan 2012, at 20:29, Robert Muir wrote:

> On Fri, Jan 27, 2012 at 3:21 PM, Hany Azzam <> wrote:
>> Hi,
>> I have two indexes. One that contains all the documents in the collection and the
other contains only the relevant documents. I am using Lucene 4.0 and the new SimilariyBase
class to build my retrieval models (similarity functions). One of the retrieval models requires
statistics to be computed across both of the indexes. How can an IndexSearcher use the two
indexes at the same time to compute different components of the retrieval model? Is that possible?
> you can make a multireader over the two indexreaders, then make an
> indexsearcher over that multireader... or are you trying to do
> something else?
> -- 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message