lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geoff Hendrey" <ghend...@decarta.com>
Subject analyzing a matched document
Date Fri, 22 May 2009 18:11:35 GMT
Let's say I have a Query that produces a set of hits in TopDocs.
 
Now, let's say the Query is a complex boolean query, but mainly it looks
at a single Field in candidate Documents.
 
Said Field may have multiple (hundreds) of values in a given Document.
For example, imaging a Field called "state" in a document representing a
Country. For the USA, the "state" field would have 50 values.
 
I would like to determine which value in the "state" Field contributed
to the score most heavily. I guess I am looking for the information
model/object that is underneath the Explanation Object. Another way to
explain my question might be like this:
 
Imagine I wanted to highlight the relevant portions of a piece of text,
that contain information that matched the query, for presentation to a
human reader. How would I find the positions in a given Field of a
Document, that were considered to be relevant to the positive search
result?
 
-geoff 
 

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