lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nicolas ThiƩbaud <ni.thieb...@gmail.com>
Subject Attributes and scoring
Date Mon, 29 Nov 2010 12:15:46 GMT
Hi everybody,

I'm a bit confused and didn't find a clearly stated answer in the
documentation.

While analyzing text, Token filters can add data, or custom data to token by
the means of Attributes (ex: TypeAttribute, OffsetAttribute, etc..). This
information is available to the following filters, but my question is, what
makes it to the index ? Do these attributes affect the scoring ?

My understanding is that scoring is only based on *tf-idf* criterion with *
coord(q,d) *(for which I didn't find documentation) and* normalization*. How
is term proximity influencing search results for example ?

Thanks for your help, any reading material is appreciated.

Nicolas.

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