You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by jreeman <mi...@hotmail.com> on 2007/03/02 11:18:16 UTC
Question about hits scores
Hello,
here is the query I am sending to lucene :
+(document_type:type0one +document_type:type0twenty0one) +state:live0
document_type and state are meta-datas. 0 is used as a separator in
"type0one " and "type0twenty0one".
I was thinking that the hits scores would be all set to 100% but it is not
the case :
I have this score :
1 hit typetype0one, score 100%
1 hit type0twenty0one, score 80%
1 hit type0twenty0one, score 80%
I am using a standart analyzer with no stop words.
I don't understand why I don't have all hits with 100% for score.
Do you have some explanation ?
Thanks in advance.
--
View this message in context: http://www.nabble.com/Question-about-hits-scores-tf3332761.html#a9267111
Sent from the Lucene - Java Developer mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
Re: [jira] Question about hits scores
Posted by jreeman <mi...@hotmail.com>.
Yes it works fine.
Thanks.
Chris Hostetter wrote:
>
>
> 1) questions about using Lucene are best sent to the java-user list ... it
> is for people asking questions about "using" the java lucene library
> (java-dev is for quesions about "developing" the java lucene library) ...
> you are more likley to get answers to questions faster on the java-user
> list.
>
> 2) calling the idf method does not modify the Similarity object ... the
> idf method is what gets called during scoring.
>
> if you wnat to change the idf method, you need to subclass
> DefaultSimilarity, override the idf method, and then pass an instance of
> your subclass to searcher.setSimilarity.
>
>
>
>
> : Date: Fri, 2 Mar 2007 04:23:49 -0800 (PST)
> : From: jreeman <mi...@hotmail.com>
> : Reply-To: java-dev@lucene.apache.org
> : To: java-dev@lucene.apache.org
> : Subject: Re: [jira] Question about hits scores
> :
> :
> : I have been overwriting the Similarity class of the searcher like this :
> :
> : DefaultSimilarity similarity = new DefaultSimilarity();
> : similarity.idf(0, 1); // idf is set to 1 whithis line of code
> : searcher.setSimilarity(similarity);
> : hits=searcher.search(l_query);
> :
> : But it does not modify the hit score.
> :
> : Do you have any suggestions ?
> :
> : jreeman wrote:
> : >
> : > Ok the score depends on the frequence of the hit for each metadata
> value.
> : >
> : > My question is now how to make the score independant from the
> frequency of
> : > a result ?
> : >
> : >
> : >
> : > jreeman wrote:
> : >>
> : >> Hello,
> : >>
> : >> here is the query I am sending to lucene :
> : >>
> : >> +(document_type:type0one +document_type:type0twenty0one)
> +state:live0
> : >>
> : >> document_type and state are meta-datas. 0 is used as a separator in
> : >> "type0one " and "type0twenty0one".
> : >>
> : >> I was thinking that the hits scores would be all set to 100% but it
> is
> : >> not the case :
> : >>
> : >> I have this score :
> : >>
> : >> 1 hit typetype0one, score 100%
> : >> 1 hit type0twenty0one, score 80%
> : >> 1 hit type0twenty0one, score 80%
> : >>
> : >> I am using a standart analyzer with no stop words.
> : >>
> : >> I don't understand why I don't have all hits with 100% for score.
> : >>
> : >> Do you have some explanation ?
> : >>
> : >> Thanks in advance.
> : >>
> : >
> : >
> :
> : --
> : View this message in context:
> http://www.nabble.com/Question-about-hits-scores-tf3332761.html#a9268663
> : Sent from the Lucene - Java Developer mailing list archive at
> Nabble.com.
> :
> :
> : ---------------------------------------------------------------------
> : To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
> : For additional commands, e-mail: java-dev-help@lucene.apache.org
> :
>
>
>
> -Hoss
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-dev-help@lucene.apache.org
>
>
>
--
View this message in context: http://www.nabble.com/Question-about-hits-scores-tf3332761.html#a9311010
Sent from the Lucene - Java Developer mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
Re: [jira] Question about hits scores
Posted by Chris Hostetter <ho...@fucit.org>.
1) questions about using Lucene are best sent to the java-user list ... it
is for people asking questions about "using" the java lucene library
(java-dev is for quesions about "developing" the java lucene library) ...
you are more likley to get answers to questions faster on the java-user
list.
2) calling the idf method does not modify the Similarity object ... the
idf method is what gets called during scoring.
if you wnat to change the idf method, you need to subclass
DefaultSimilarity, override the idf method, and then pass an instance of
your subclass to searcher.setSimilarity.
: Date: Fri, 2 Mar 2007 04:23:49 -0800 (PST)
: From: jreeman <mi...@hotmail.com>
: Reply-To: java-dev@lucene.apache.org
: To: java-dev@lucene.apache.org
: Subject: Re: [jira] Question about hits scores
:
:
: I have been overwriting the Similarity class of the searcher like this :
:
: DefaultSimilarity similarity = new DefaultSimilarity();
: similarity.idf(0, 1); // idf is set to 1 whithis line of code
: searcher.setSimilarity(similarity);
: hits=searcher.search(l_query);
:
: But it does not modify the hit score.
:
: Do you have any suggestions ?
:
: jreeman wrote:
: >
: > Ok the score depends on the frequence of the hit for each metadata value.
: >
: > My question is now how to make the score independant from the frequency of
: > a result ?
: >
: >
: >
: > jreeman wrote:
: >>
: >> Hello,
: >>
: >> here is the query I am sending to lucene :
: >>
: >> +(document_type:type0one +document_type:type0twenty0one) +state:live0
: >>
: >> document_type and state are meta-datas. 0 is used as a separator in
: >> "type0one " and "type0twenty0one".
: >>
: >> I was thinking that the hits scores would be all set to 100% but it is
: >> not the case :
: >>
: >> I have this score :
: >>
: >> 1 hit typetype0one, score 100%
: >> 1 hit type0twenty0one, score 80%
: >> 1 hit type0twenty0one, score 80%
: >>
: >> I am using a standart analyzer with no stop words.
: >>
: >> I don't understand why I don't have all hits with 100% for score.
: >>
: >> Do you have some explanation ?
: >>
: >> Thanks in advance.
: >>
: >
: >
:
: --
: View this message in context: http://www.nabble.com/Question-about-hits-scores-tf3332761.html#a9268663
: Sent from the Lucene - Java Developer mailing list archive at Nabble.com.
:
:
: ---------------------------------------------------------------------
: To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
: For additional commands, e-mail: java-dev-help@lucene.apache.org
:
-Hoss
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
Re: [jira] Question about hits scores
Posted by jreeman <mi...@hotmail.com>.
I have been overwriting the Similarity class of the searcher like this :
DefaultSimilarity similarity = new DefaultSimilarity();
similarity.idf(0, 1); // idf is set to 1 whithis line of code
searcher.setSimilarity(similarity);
hits=searcher.search(l_query);
But it does not modify the hit score.
Do you have any suggestions ?
jreeman wrote:
>
> Ok the score depends on the frequence of the hit for each metadata value.
>
> My question is now how to make the score independant from the frequency of
> a result ?
>
>
>
> jreeman wrote:
>>
>> Hello,
>>
>> here is the query I am sending to lucene :
>>
>> +(document_type:type0one +document_type:type0twenty0one) +state:live0
>>
>> document_type and state are meta-datas. 0 is used as a separator in
>> "type0one " and "type0twenty0one".
>>
>> I was thinking that the hits scores would be all set to 100% but it is
>> not the case :
>>
>> I have this score :
>>
>> 1 hit typetype0one, score 100%
>> 1 hit type0twenty0one, score 80%
>> 1 hit type0twenty0one, score 80%
>>
>> I am using a standart analyzer with no stop words.
>>
>> I don't understand why I don't have all hits with 100% for score.
>>
>> Do you have some explanation ?
>>
>> Thanks in advance.
>>
>
>
--
View this message in context: http://www.nabble.com/Question-about-hits-scores-tf3332761.html#a9268663
Sent from the Lucene - Java Developer mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
Re: [jira] Question about hits scores
Posted by jreeman <mi...@hotmail.com>.
Ok the score depends on the frequence of the hit for each metadata value.
My question is now how to make the score independant from the frequency of a
result ?
jreeman wrote:
>
> Hello,
>
> here is the query I am sending to lucene :
>
> +(document_type:type0one +document_type:type0twenty0one) +state:live0
>
> document_type and state are meta-datas. 0 is used as a separator in
> "type0one " and "type0twenty0one".
>
> I was thinking that the hits scores would be all set to 100% but it is not
> the case :
>
> I have this score :
>
> 1 hit typetype0one, score 100%
> 1 hit type0twenty0one, score 80%
> 1 hit type0twenty0one, score 80%
>
> I am using a standart analyzer with no stop words.
>
> I don't understand why I don't have all hits with 100% for score.
>
> Do you have some explanation ?
>
> Thanks in advance.
>
--
View this message in context: http://www.nabble.com/Question-about-hits-scores-tf3332761.html#a9267212
Sent from the Lucene - Java Developer mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org