You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by nishi <ni...@gmail.com> on 2013/05/06 13:33:08 UTC
Scores dilemma after providing boosting with bq as same weigtage
for 2 condition
While giving bq with same weightage as ^1.2 below on two values of
articleTopic, the result always coming all of the "Office" on the top
somehow. What other factors would influence at this scenario when there is
no keyword search also?
http://localhost:8080/solr?rows=900&fq=(articleTopic:"Food" OR
articleTopic:"Office")&bq=(articleTopic:"Food" OR
articleTopic:"Office")^1.2&fl=title,description,documentId,score&sort=score
desc
Also with other fields adding at bq, still the results favoured somehow to
"Office" one:
http://localhost:8080/solr?rows=900&fq=(articleTopic:"Food" OR
articleTopic:"Office")&bq=(articleSrc:"News" OR articleSrc:"Blog")^0.5 OR
(articleTopic:"Food" OR
articleTopic:"Office")^1.2&fl=title,description,documentId,score&sort=score
desc
Please advice what might be the reason/factors which doesn't balance the
result with both articleTopic and instead always favored the result with the
"Food" somehow in the score.
Thanks in advance
--
View this message in context: http://lucene.472066.n3.nabble.com/Scores-dilemma-after-providing-boosting-with-bq-as-same-weigtage-for-2-condition-tp4061035.html
Sent from the Solr - User mailing list archive at Nabble.com.
Re: Scores dilemma after providing boosting with bq as same
weigtage for 2 condition
Posted by nishi <ni...@gmail.com>.
"ab_1eb83ef9bc0896":"
0.17063755 = (MATCH) sum of:
3.9999085E-4 = (MATCH) MatchAllDocsQuery, product of:
3.9999085E-4 = queryNorm
0.009742409 = (MATCH) product of:
0.019484818 = (MATCH) sum of:
0.016588148 = (MATCH) sum of:
0.0034696688 = (MATCH) weight(articleTopic:Food^1.2 in 2441)
[DefaultSimilarity], result of:
0.0034696688 = score(doc=2441,freq=1.0 = termFreq=1.0
), product of:
0.0012905049 = queryWeight, product of:
1.2 = boost
2.6886134 = idf(docFreq=52556, maxDocs=284437)
3.9999085E-4 = queryNorm
2.6886134 = fieldWeight in 2441, product of:
1.0 = tf(freq=1.0), with freq of:
1.0 = termFreq=1.0
2.6886134 = idf(docFreq=52556, maxDocs=284437)
1.0 = fieldNorm(doc=2441)
0.013118479 = (MATCH) weight(articleTopic:Office^1.2 in 2441)
[DefaultSimilarity], result of:
0.013118479 = score(doc=2441,freq=1.0 = termFreq=1.0
), product of:
0.0025093278 = queryWeight, product of:
1.2 = boost
5.2278857 = idf(docFreq=4147, maxDocs=284437)
3.9999085E-4 = queryNorm
5.2278857 = fieldWeight in 2441, product of:
1.0 = tf(freq=1.0), with freq of:
1.0 = termFreq=1.0
5.2278857 = idf(docFreq=4147, maxDocs=284437)
1.0 = fieldNorm(doc=2441)
7.967604E-4 = (MATCH) product of:
0.0051789423 = (MATCH) sum of:
0.0017619515 = (MATCH) weight(subTopic:Protein in 2441)
[DefaultSimilarity], result of:
0.0017619515 = score(doc=2441,freq=1.0 = termFreq=1.0
), product of:
4.5981447E-4 = queryWeight, product of:
3.8318748 = idf(docFreq=16753, maxDocs=284437)
1.1999726E-4 = queryNorm
3.8318748 = fieldWeight in 2441, product of:
1.0 = tf(freq=1.0), with freq of:
1.0 = termFreq=1.0
3.8318748 = idf(docFreq=16753, maxDocs=284437)
1.0 = fieldNorm(doc=2441)
0.0034169909 = (MATCH) weight(subTopic:Printers in 2441)
[DefaultSimilarity], result of:
0.0034169909 = score(doc=2441,freq=1.0 = termFreq=1.0
), product of:
5.019797E-4 = queryWeight, product of:
0.3 = boost
4.18326 = idf(docFreq=11789, maxDocs=284437)
3.9999085E-4 = queryNorm
4.18326 = fieldWeight in 2441, product of:
1.0 = tf(freq=1.0), with freq of:
1.0 = termFreq=1.0
4.18326 = idf(docFreq=11789, maxDocs=284437)
1.0 = fieldNorm(doc=2441)
0.5 = coord(3/6)
0.16049515 = (MATCH)
FunctionQuery(0.08/(3.16E-11*float(ms(const(1367798400000),date(pubDate)))+0.5)),
product of:
0.16049883 =
0.08/(3.16E-11*float(ms(const(1367798400000),date(pubDate)=1367847578000))+0.5)
2500.0 = boost
3.9999085E-4 = queryNorm
",
.
.
.
..................
This is the explain description for the score coming up....but not coming in
easier understandable format...any pointer would be helpful, meantime
looking into it to understand more
--
View this message in context: http://lucene.472066.n3.nabble.com/Scores-dilemma-after-providing-boosting-with-bq-as-same-weigtage-for-2-condition-tp4061035p4061480.html
Sent from the Solr - User mailing list archive at Nabble.com.
Re: Scores dilemma after providing boosting with bq as same weigtage
for 2 condition
Posted by Erick Erickson <er...@gmail.com>.
Try adding &debugQuery=true to your query, the resulting
data will show you exactly how the doc score is calculated.
Warning: reading the explain can be a bit challenging, but
that's the only way to really understand why docs scored as
they did.
Best
Erick
On Mon, May 6, 2013 at 7:33 AM, nishi <ni...@gmail.com> wrote:
> While giving bq with same weightage as ^1.2 below on two values of
> articleTopic, the result always coming all of the "Office" on the top
> somehow. What other factors would influence at this scenario when there is
> no keyword search also?
>
> http://localhost:8080/solr?rows=900&fq=(articleTopic:"Food" OR
> articleTopic:"Office")&bq=(articleTopic:"Food" OR
> articleTopic:"Office")^1.2&fl=title,description,documentId,score&sort=score
> desc
>
> Also with other fields adding at bq, still the results favoured somehow to
> "Office" one:
> http://localhost:8080/solr?rows=900&fq=(articleTopic:"Food" OR
> articleTopic:"Office")&bq=(articleSrc:"News" OR articleSrc:"Blog")^0.5 OR
> (articleTopic:"Food" OR
> articleTopic:"Office")^1.2&fl=title,description,documentId,score&sort=score
> desc
>
> Please advice what might be the reason/factors which doesn't balance the
> result with both articleTopic and instead always favored the result with the
> "Food" somehow in the score.
>
> Thanks in advance
>
>
>
>
> --
> View this message in context: http://lucene.472066.n3.nabble.com/Scores-dilemma-after-providing-boosting-with-bq-as-same-weigtage-for-2-condition-tp4061035.html
> Sent from the Solr - User mailing list archive at Nabble.com.