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 Sravan Kumar <sr...@caavo.com> on 2018/01/31 12:38:45 UTC
Title Search scoring issues with multivalued field & norm
Hi,
We are using solr for our movie title search.
As it is "title search", this should be treated different than the normal
document search.
Hence, we use a modified version of TFIDFSimilarity with the following
changes.
- disabled TF & IDF and will only have 1 as value.
- disabled norms by specifying omitNorms as true for all the fields.
There are 6 fields with different analyzers and we make use of different
weights in edismax's qf & pf parameters to match tokens & boost phrases.
But, movies could have aliases and have multiple titles. So, we made the
fields multivalued.
Now, consider the following four documents
1> "Beauty and the Beast"
2> "The Real Beauty and the Beast"
3> "Beauty and the Beast", "La bella y la bestia"
4> "Beauty and the Beast"
Note: Document 3 has two titles in it.
So, for a query "Beauty and the Beast" and with the above configuration all
the documents receive same score. But 1,3,4 should have got same score and
document 2 lesser than others.
To solve this, we followed what is suggested in the following thread:
http://lucene.472066.n3.nabble.com/Influencing-scores-on-values-in-multiValue-fields-td1791651.html
Now, the fields which are used to boost are made to use Norms. And for
matching norms are disabled. This is to make sure that exact & near exact
matches are rewarded.
But, for the same query, we get the following results.
query: "Beauty & the Beast"
Search Results:
1> "Beauty and the Beast"
4> "Beauty and the Beast"
2> "The Real Beauty and the Beast"
3> "Beauty and the Beast", "La bella y la bestia"
Clearly, the changes have solved only a part of the problem. The document 3
should be ranked/scored higher than document 2.
This is because lucene considers the total field length across all the
values in a multivalued field for normalization.
How do we handle this scenario and make sure that in multivalued fields the
normalization is taken care of?
--
Regards,
Sravan
Re: Title Search scoring issues with multivalued field & norm
Posted by Sravan Kumar <sr...@caavo.com>.
Using edismax with different fields for each title will affect the final
scores if the tie paramter is non-zero.
Can we create separate document for each title? The uniqueness won't be for
movie_id but for each title. In this manner, even while using edismax, the
other titles won't affect the score.
Any other way to handle norms in multivalued field?
On Thu, Feb 1, 2018 at 12:24 PM, Sravan Kumar <sr...@caavo.com> wrote:
> @Walter: Perhaps you are right on not to consider stemming. Instead fuzzy
> search will cover these along with the misspellings.
>
> In case of symbols, we want the titles matching the symbols ranked higher
> than the others. Perhaps we can use this field only for boosting.
>
> Certain movies have around 4-6 different aliases based on what our source
> gives and we do not really know what is the max. Is there no other way from
> lucene/solr to use a multivalued field?
>
>
> On Thu, Feb 1, 2018 at 11:06 AM, Walter Underwood <wu...@wunderwood.org>
> wrote:
>
>> I was the first search engineer at Netflix and moved their search from a
>> home-grown engine to Solr. It worked very well with a single title field
>> and aliases.
>>
>> I think your schema is too complicated for movie search.
>>
>> Stemming is not useful. It doesn’t help search and it can hurt. You don’t
>> want the movie “Saw” to match the query “see”.
>>
>> When is it useful to search with symbols? Remove the punctuation.
>>
>> The only movie titles with symbols that caused any challenge were:
>>
>> * Frost/Nixon
>> * .hack//Sign
>> * +/-
>>
>> For the first two, removing punctuation worked fine. For the last one, I
>> hardcoded a translation to “plus/minus” before indexing or querying.
>>
>> Query completion made a huge difference, taking our clickthrough rate
>> from 0.45 to 0.55.
>>
>> Later, we added fuzzy search to handle misspellings.
>>
>> wunder
>> Walter Underwood
>> wunder@wunderwood.org
>> http://observer.wunderwood.org/ (my blog)
>>
>> > On Jan 31, 2018, at 8:54 PM, Sravan Kumar <sr...@caavo.com> wrote:
>> >
>> > @Tim Casey: Yeah... TFIDFSimilarity weighs towards shorter documents.
>> This
>> > is done through the fieldnorm component in the class. The issue is when
>> the
>> > field is multivalued. Consider the field has two string each of 4
>> tokens.
>> > The fieldNorm from the lucene TFIDFSimilarity class considers the total
>> sum
>> > of these two values i.e 8 for normalizing instead of 4. Hence, the
>> ranking
>> > is distorted.
>> > Regarding the search evaluation, we do have a curated set.
>> >
>> >
>> > On Thu, Feb 1, 2018 at 9:18 AM, Tim Casey <tc...@gmail.com> wrote:
>> >
>> >> For smaller length documents TFIDFSimilarity will weight towards
>> shorter
>> >> documents. Another way to say this, if your documents are 5-10 terms,
>> the
>> >> 5 terms are going to win.
>> >> You might think about having per token, or token pair, weight. I
>> would be
>> >> surprised if there was not something similar out there. This is a
>> common
>> >> issue with any short text.
>> >> I guess I would think of this as TFICF, where the CF is the corpus
>> >> frequency. You also might want to weight inversely proportional to the
>> age
>> >> of the title, older are less important. This is assuming people are
>> doing
>> >> searches within some time cluster, newer is more likely.
>> >>
>> >> For some obvious advice, things you probably already know. This kind
>> of
>> >> search needs some hard measurement to begin to know how to tune it.
>> You
>> >> need to find a reasonable annotated representation. So, if you took
>> the
>> >> previous months searches where there is a chain of successive
>> searches. If
>> >> you weighted things differently would you shorten the length of the
>> chain.
>> >> Can you get the click throughs to happen sooner.
>> >>
>> >> Anyway, just my 2 cents....
>> >>
>> >>
>> >> On Wed, Jan 31, 2018 at 6:38 PM, Sravan Kumar <sr...@caavo.com>
>> wrote:
>> >>
>> >>>
>> >>> @Walter: We have 6 fields declared in schema.xml for title each with
>> >>> different type of analyzer. One without processing symbols, other
>> stemmed
>> >>> and other removing symbols, etc. So, if we have separate fields for
>> each
>> >>> alias it will be that many times the number of final fields declared
>> in
>> >>> schema.xml. And we exactly do not know what is the maximum number of
>> >>> aliases a movie can have.
>> >>> @Walter: I will try this but isn’t there any other way where I can
>> >> tweak ?
>> >>>
>> >>> @eric: will try this. But it will work only for exact matches.
>> >>>
>> >>>
>> >>>> On Jan 31, 2018, at 10:39 PM, Erick Erickson <
>> erickerickson@gmail.com>
>> >>> wrote:
>> >>>>
>> >>>> Or use a boost for the phrase, something like
>> >>>> "beauty and the beast"^5
>> >>>>
>> >>>>> On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <
>> >>> wunder@wunderwood.org> wrote:
>> >>>>> You can use a separate field for title aliases. That is what I did
>> for
>> >>> Netflix search.
>> >>>>>
>> >>>>> Why disable idf? Disabling tf for titles can be a good idea, for
>> >>> example the movie “New York, New York” is not twice as much about New
>> >> York
>> >>> as some other film that just lists it once.
>> >>>>>
>> >>>>> Also, consider using a popularity score as a boost.
>> >>>>>
>> >>>>> wunder
>> >>>>> Walter Underwood
>> >>>>> wunder@wunderwood.org
>> >>>>> http://observer.wunderwood.org/ (my blog)
>> >>>>>
>> >>>>>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com>
>> wrote:
>> >>>>>>
>> >>>>>> Hi,
>> >>>>>> We are using solr for our movie title search.
>> >>>>>>
>> >>>>>>
>> >>>>>> As it is "title search", this should be treated different than the
>> >>> normal
>> >>>>>> document search.
>> >>>>>> Hence, we use a modified version of TFIDFSimilarity with the
>> >> following
>> >>>>>> changes.
>> >>>>>> - disabled TF & IDF and will only have 1 as value.
>> >>>>>> - disabled norms by specifying omitNorms as true for all the
>> fields.
>> >>>>>>
>> >>>>>> There are 6 fields with different analyzers and we make use of
>> >>> different
>> >>>>>> weights in edismax's qf & pf parameters to match tokens & boost
>> >>> phrases.
>> >>>>>>
>> >>>>>> But, movies could have aliases and have multiple titles. So, we
>> made
>> >>> the
>> >>>>>> fields multivalued.
>> >>>>>>
>> >>>>>> Now, consider the following four documents
>> >>>>>> 1> "Beauty and the Beast"
>> >>>>>> 2> "The Real Beauty and the Beast"
>> >>>>>> 3> "Beauty and the Beast", "La bella y la bestia"
>> >>>>>> 4> "Beauty and the Beast"
>> >>>>>>
>> >>>>>> Note: Document 3 has two titles in it.
>> >>>>>>
>> >>>>>> So, for a query "Beauty and the Beast" and with the above
>> >>> configuration all
>> >>>>>> the documents receive same score. But 1,3,4 should have got same
>> >> score
>> >>> and
>> >>>>>> document 2 lesser than others.
>> >>>>>>
>> >>>>>> To solve this, we followed what is suggested in the following
>> thread:
>> >>>>>> http://lucene.472066.n3.nabble.com/Influencing-scores-
>> >>> on-values-in-multiValue-fields-td1791651.html
>> >>>>>>
>> >>>>>> Now, the fields which are used to boost are made to use Norms. And
>> >> for
>> >>>>>> matching norms are disabled. This is to make sure that exact & near
>> >>> exact
>> >>>>>> matches are rewarded.
>> >>>>>>
>> >>>>>> But, for the same query, we get the following results.
>> >>>>>> query: "Beauty & the Beast"
>> >>>>>> Search Results:
>> >>>>>> 1> "Beauty and the Beast"
>> >>>>>> 4> "Beauty and the Beast"
>> >>>>>> 2> "The Real Beauty and the Beast"
>> >>>>>> 3> "Beauty and the Beast", "La bella y la bestia"
>> >>>>>>
>> >>>>>> Clearly, the changes have solved only a part of the problem. The
>> >>> document 3
>> >>>>>> should be ranked/scored higher than document 2.
>> >>>>>>
>> >>>>>> This is because lucene considers the total field length across all
>> >> the
>> >>>>>> values in a multivalued field for normalization.
>> >>>>>>
>> >>>>>> How do we handle this scenario and make sure that in multivalued
>> >>> fields the
>> >>>>>> normalization is taken care of?
>> >>>>>>
>> >>>>>>
>> >>>>>> --
>> >>>>>> Regards,
>> >>>>>> Sravan
>> >>>>>
>> >>>
>> >>
>> >
>> >
>> >
>> > --
>> > Regards,
>> > Sravan
>>
>>
>
>
> --
> Regards,
> Sravan
>
--
Regards,
Sravan
Re: Title Search scoring issues with multivalued field & norm
Posted by Sravan Kumar <sr...@caavo.com>.
@Walter: Perhaps you are right on not to consider stemming. Instead fuzzy
search will cover these along with the misspellings.
In case of symbols, we want the titles matching the symbols ranked higher
than the others. Perhaps we can use this field only for boosting.
Certain movies have around 4-6 different aliases based on what our source
gives and we do not really know what is the max. Is there no other way from
lucene/solr to use a multivalued field?
On Thu, Feb 1, 2018 at 11:06 AM, Walter Underwood <wu...@wunderwood.org>
wrote:
> I was the first search engineer at Netflix and moved their search from a
> home-grown engine to Solr. It worked very well with a single title field
> and aliases.
>
> I think your schema is too complicated for movie search.
>
> Stemming is not useful. It doesn’t help search and it can hurt. You don’t
> want the movie “Saw” to match the query “see”.
>
> When is it useful to search with symbols? Remove the punctuation.
>
> The only movie titles with symbols that caused any challenge were:
>
> * Frost/Nixon
> * .hack//Sign
> * +/-
>
> For the first two, removing punctuation worked fine. For the last one, I
> hardcoded a translation to “plus/minus” before indexing or querying.
>
> Query completion made a huge difference, taking our clickthrough rate from
> 0.45 to 0.55.
>
> Later, we added fuzzy search to handle misspellings.
>
> wunder
> Walter Underwood
> wunder@wunderwood.org
> http://observer.wunderwood.org/ (my blog)
>
> > On Jan 31, 2018, at 8:54 PM, Sravan Kumar <sr...@caavo.com> wrote:
> >
> > @Tim Casey: Yeah... TFIDFSimilarity weighs towards shorter documents.
> This
> > is done through the fieldnorm component in the class. The issue is when
> the
> > field is multivalued. Consider the field has two string each of 4 tokens.
> > The fieldNorm from the lucene TFIDFSimilarity class considers the total
> sum
> > of these two values i.e 8 for normalizing instead of 4. Hence, the
> ranking
> > is distorted.
> > Regarding the search evaluation, we do have a curated set.
> >
> >
> > On Thu, Feb 1, 2018 at 9:18 AM, Tim Casey <tc...@gmail.com> wrote:
> >
> >> For smaller length documents TFIDFSimilarity will weight towards shorter
> >> documents. Another way to say this, if your documents are 5-10 terms,
> the
> >> 5 terms are going to win.
> >> You might think about having per token, or token pair, weight. I would
> be
> >> surprised if there was not something similar out there. This is a
> common
> >> issue with any short text.
> >> I guess I would think of this as TFICF, where the CF is the corpus
> >> frequency. You also might want to weight inversely proportional to the
> age
> >> of the title, older are less important. This is assuming people are
> doing
> >> searches within some time cluster, newer is more likely.
> >>
> >> For some obvious advice, things you probably already know. This kind of
> >> search needs some hard measurement to begin to know how to tune it. You
> >> need to find a reasonable annotated representation. So, if you took the
> >> previous months searches where there is a chain of successive
> searches. If
> >> you weighted things differently would you shorten the length of the
> chain.
> >> Can you get the click throughs to happen sooner.
> >>
> >> Anyway, just my 2 cents....
> >>
> >>
> >> On Wed, Jan 31, 2018 at 6:38 PM, Sravan Kumar <sr...@caavo.com> wrote:
> >>
> >>>
> >>> @Walter: We have 6 fields declared in schema.xml for title each with
> >>> different type of analyzer. One without processing symbols, other
> stemmed
> >>> and other removing symbols, etc. So, if we have separate fields for
> each
> >>> alias it will be that many times the number of final fields declared in
> >>> schema.xml. And we exactly do not know what is the maximum number of
> >>> aliases a movie can have.
> >>> @Walter: I will try this but isn’t there any other way where I can
> >> tweak ?
> >>>
> >>> @eric: will try this. But it will work only for exact matches.
> >>>
> >>>
> >>>> On Jan 31, 2018, at 10:39 PM, Erick Erickson <erickerickson@gmail.com
> >
> >>> wrote:
> >>>>
> >>>> Or use a boost for the phrase, something like
> >>>> "beauty and the beast"^5
> >>>>
> >>>>> On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <
> >>> wunder@wunderwood.org> wrote:
> >>>>> You can use a separate field for title aliases. That is what I did
> for
> >>> Netflix search.
> >>>>>
> >>>>> Why disable idf? Disabling tf for titles can be a good idea, for
> >>> example the movie “New York, New York” is not twice as much about New
> >> York
> >>> as some other film that just lists it once.
> >>>>>
> >>>>> Also, consider using a popularity score as a boost.
> >>>>>
> >>>>> wunder
> >>>>> Walter Underwood
> >>>>> wunder@wunderwood.org
> >>>>> http://observer.wunderwood.org/ (my blog)
> >>>>>
> >>>>>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
> >>>>>>
> >>>>>> Hi,
> >>>>>> We are using solr for our movie title search.
> >>>>>>
> >>>>>>
> >>>>>> As it is "title search", this should be treated different than the
> >>> normal
> >>>>>> document search.
> >>>>>> Hence, we use a modified version of TFIDFSimilarity with the
> >> following
> >>>>>> changes.
> >>>>>> - disabled TF & IDF and will only have 1 as value.
> >>>>>> - disabled norms by specifying omitNorms as true for all the
> fields.
> >>>>>>
> >>>>>> There are 6 fields with different analyzers and we make use of
> >>> different
> >>>>>> weights in edismax's qf & pf parameters to match tokens & boost
> >>> phrases.
> >>>>>>
> >>>>>> But, movies could have aliases and have multiple titles. So, we made
> >>> the
> >>>>>> fields multivalued.
> >>>>>>
> >>>>>> Now, consider the following four documents
> >>>>>> 1> "Beauty and the Beast"
> >>>>>> 2> "The Real Beauty and the Beast"
> >>>>>> 3> "Beauty and the Beast", "La bella y la bestia"
> >>>>>> 4> "Beauty and the Beast"
> >>>>>>
> >>>>>> Note: Document 3 has two titles in it.
> >>>>>>
> >>>>>> So, for a query "Beauty and the Beast" and with the above
> >>> configuration all
> >>>>>> the documents receive same score. But 1,3,4 should have got same
> >> score
> >>> and
> >>>>>> document 2 lesser than others.
> >>>>>>
> >>>>>> To solve this, we followed what is suggested in the following
> thread:
> >>>>>> http://lucene.472066.n3.nabble.com/Influencing-scores-
> >>> on-values-in-multiValue-fields-td1791651.html
> >>>>>>
> >>>>>> Now, the fields which are used to boost are made to use Norms. And
> >> for
> >>>>>> matching norms are disabled. This is to make sure that exact & near
> >>> exact
> >>>>>> matches are rewarded.
> >>>>>>
> >>>>>> But, for the same query, we get the following results.
> >>>>>> query: "Beauty & the Beast"
> >>>>>> Search Results:
> >>>>>> 1> "Beauty and the Beast"
> >>>>>> 4> "Beauty and the Beast"
> >>>>>> 2> "The Real Beauty and the Beast"
> >>>>>> 3> "Beauty and the Beast", "La bella y la bestia"
> >>>>>>
> >>>>>> Clearly, the changes have solved only a part of the problem. The
> >>> document 3
> >>>>>> should be ranked/scored higher than document 2.
> >>>>>>
> >>>>>> This is because lucene considers the total field length across all
> >> the
> >>>>>> values in a multivalued field for normalization.
> >>>>>>
> >>>>>> How do we handle this scenario and make sure that in multivalued
> >>> fields the
> >>>>>> normalization is taken care of?
> >>>>>>
> >>>>>>
> >>>>>> --
> >>>>>> Regards,
> >>>>>> Sravan
> >>>>>
> >>>
> >>
> >
> >
> >
> > --
> > Regards,
> > Sravan
>
>
--
Regards,
Sravan
Re: Title Search scoring issues with multivalued field & norm
Posted by Walter Underwood <wu...@wunderwood.org>.
I was the first search engineer at Netflix and moved their search from a home-grown engine to Solr. It worked very well with a single title field and aliases.
I think your schema is too complicated for movie search.
Stemming is not useful. It doesn’t help search and it can hurt. You don’t want the movie “Saw” to match the query “see”.
When is it useful to search with symbols? Remove the punctuation.
The only movie titles with symbols that caused any challenge were:
* Frost/Nixon
* .hack//Sign
* +/-
For the first two, removing punctuation worked fine. For the last one, I hardcoded a translation to “plus/minus” before indexing or querying.
Query completion made a huge difference, taking our clickthrough rate from 0.45 to 0.55.
Later, we added fuzzy search to handle misspellings.
wunder
Walter Underwood
wunder@wunderwood.org
http://observer.wunderwood.org/ (my blog)
> On Jan 31, 2018, at 8:54 PM, Sravan Kumar <sr...@caavo.com> wrote:
>
> @Tim Casey: Yeah... TFIDFSimilarity weighs towards shorter documents. This
> is done through the fieldnorm component in the class. The issue is when the
> field is multivalued. Consider the field has two string each of 4 tokens.
> The fieldNorm from the lucene TFIDFSimilarity class considers the total sum
> of these two values i.e 8 for normalizing instead of 4. Hence, the ranking
> is distorted.
> Regarding the search evaluation, we do have a curated set.
>
>
> On Thu, Feb 1, 2018 at 9:18 AM, Tim Casey <tc...@gmail.com> wrote:
>
>> For smaller length documents TFIDFSimilarity will weight towards shorter
>> documents. Another way to say this, if your documents are 5-10 terms, the
>> 5 terms are going to win.
>> You might think about having per token, or token pair, weight. I would be
>> surprised if there was not something similar out there. This is a common
>> issue with any short text.
>> I guess I would think of this as TFICF, where the CF is the corpus
>> frequency. You also might want to weight inversely proportional to the age
>> of the title, older are less important. This is assuming people are doing
>> searches within some time cluster, newer is more likely.
>>
>> For some obvious advice, things you probably already know. This kind of
>> search needs some hard measurement to begin to know how to tune it. You
>> need to find a reasonable annotated representation. So, if you took the
>> previous months searches where there is a chain of successive searches. If
>> you weighted things differently would you shorten the length of the chain.
>> Can you get the click throughs to happen sooner.
>>
>> Anyway, just my 2 cents....
>>
>>
>> On Wed, Jan 31, 2018 at 6:38 PM, Sravan Kumar <sr...@caavo.com> wrote:
>>
>>>
>>> @Walter: We have 6 fields declared in schema.xml for title each with
>>> different type of analyzer. One without processing symbols, other stemmed
>>> and other removing symbols, etc. So, if we have separate fields for each
>>> alias it will be that many times the number of final fields declared in
>>> schema.xml. And we exactly do not know what is the maximum number of
>>> aliases a movie can have.
>>> @Walter: I will try this but isn’t there any other way where I can
>> tweak ?
>>>
>>> @eric: will try this. But it will work only for exact matches.
>>>
>>>
>>>> On Jan 31, 2018, at 10:39 PM, Erick Erickson <er...@gmail.com>
>>> wrote:
>>>>
>>>> Or use a boost for the phrase, something like
>>>> "beauty and the beast"^5
>>>>
>>>>> On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <
>>> wunder@wunderwood.org> wrote:
>>>>> You can use a separate field for title aliases. That is what I did for
>>> Netflix search.
>>>>>
>>>>> Why disable idf? Disabling tf for titles can be a good idea, for
>>> example the movie “New York, New York” is not twice as much about New
>> York
>>> as some other film that just lists it once.
>>>>>
>>>>> Also, consider using a popularity score as a boost.
>>>>>
>>>>> wunder
>>>>> Walter Underwood
>>>>> wunder@wunderwood.org
>>>>> http://observer.wunderwood.org/ (my blog)
>>>>>
>>>>>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
>>>>>>
>>>>>> Hi,
>>>>>> We are using solr for our movie title search.
>>>>>>
>>>>>>
>>>>>> As it is "title search", this should be treated different than the
>>> normal
>>>>>> document search.
>>>>>> Hence, we use a modified version of TFIDFSimilarity with the
>> following
>>>>>> changes.
>>>>>> - disabled TF & IDF and will only have 1 as value.
>>>>>> - disabled norms by specifying omitNorms as true for all the fields.
>>>>>>
>>>>>> There are 6 fields with different analyzers and we make use of
>>> different
>>>>>> weights in edismax's qf & pf parameters to match tokens & boost
>>> phrases.
>>>>>>
>>>>>> But, movies could have aliases and have multiple titles. So, we made
>>> the
>>>>>> fields multivalued.
>>>>>>
>>>>>> Now, consider the following four documents
>>>>>> 1> "Beauty and the Beast"
>>>>>> 2> "The Real Beauty and the Beast"
>>>>>> 3> "Beauty and the Beast", "La bella y la bestia"
>>>>>> 4> "Beauty and the Beast"
>>>>>>
>>>>>> Note: Document 3 has two titles in it.
>>>>>>
>>>>>> So, for a query "Beauty and the Beast" and with the above
>>> configuration all
>>>>>> the documents receive same score. But 1,3,4 should have got same
>> score
>>> and
>>>>>> document 2 lesser than others.
>>>>>>
>>>>>> To solve this, we followed what is suggested in the following thread:
>>>>>> http://lucene.472066.n3.nabble.com/Influencing-scores-
>>> on-values-in-multiValue-fields-td1791651.html
>>>>>>
>>>>>> Now, the fields which are used to boost are made to use Norms. And
>> for
>>>>>> matching norms are disabled. This is to make sure that exact & near
>>> exact
>>>>>> matches are rewarded.
>>>>>>
>>>>>> But, for the same query, we get the following results.
>>>>>> query: "Beauty & the Beast"
>>>>>> Search Results:
>>>>>> 1> "Beauty and the Beast"
>>>>>> 4> "Beauty and the Beast"
>>>>>> 2> "The Real Beauty and the Beast"
>>>>>> 3> "Beauty and the Beast", "La bella y la bestia"
>>>>>>
>>>>>> Clearly, the changes have solved only a part of the problem. The
>>> document 3
>>>>>> should be ranked/scored higher than document 2.
>>>>>>
>>>>>> This is because lucene considers the total field length across all
>> the
>>>>>> values in a multivalued field for normalization.
>>>>>>
>>>>>> How do we handle this scenario and make sure that in multivalued
>>> fields the
>>>>>> normalization is taken care of?
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Regards,
>>>>>> Sravan
>>>>>
>>>
>>
>
>
>
> --
> Regards,
> Sravan
Re: Title Search scoring issues with multivalued field & norm
Posted by Sravan Kumar <sr...@caavo.com>.
@Tim Casey: Yeah... TFIDFSimilarity weighs towards shorter documents. This
is done through the fieldnorm component in the class. The issue is when the
field is multivalued. Consider the field has two string each of 4 tokens.
The fieldNorm from the lucene TFIDFSimilarity class considers the total sum
of these two values i.e 8 for normalizing instead of 4. Hence, the ranking
is distorted.
Regarding the search evaluation, we do have a curated set.
On Thu, Feb 1, 2018 at 9:18 AM, Tim Casey <tc...@gmail.com> wrote:
> For smaller length documents TFIDFSimilarity will weight towards shorter
> documents. Another way to say this, if your documents are 5-10 terms, the
> 5 terms are going to win.
> You might think about having per token, or token pair, weight. I would be
> surprised if there was not something similar out there. This is a common
> issue with any short text.
> I guess I would think of this as TFICF, where the CF is the corpus
> frequency. You also might want to weight inversely proportional to the age
> of the title, older are less important. This is assuming people are doing
> searches within some time cluster, newer is more likely.
>
> For some obvious advice, things you probably already know. This kind of
> search needs some hard measurement to begin to know how to tune it. You
> need to find a reasonable annotated representation. So, if you took the
> previous months searches where there is a chain of successive searches. If
> you weighted things differently would you shorten the length of the chain.
> Can you get the click throughs to happen sooner.
>
> Anyway, just my 2 cents....
>
>
> On Wed, Jan 31, 2018 at 6:38 PM, Sravan Kumar <sr...@caavo.com> wrote:
>
> >
> > @Walter: We have 6 fields declared in schema.xml for title each with
> > different type of analyzer. One without processing symbols, other stemmed
> > and other removing symbols, etc. So, if we have separate fields for each
> > alias it will be that many times the number of final fields declared in
> > schema.xml. And we exactly do not know what is the maximum number of
> > aliases a movie can have.
> > @Walter: I will try this but isn’t there any other way where I can
> tweak ?
> >
> > @eric: will try this. But it will work only for exact matches.
> >
> >
> > > On Jan 31, 2018, at 10:39 PM, Erick Erickson <er...@gmail.com>
> > wrote:
> > >
> > > Or use a boost for the phrase, something like
> > > "beauty and the beast"^5
> > >
> > >> On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <
> > wunder@wunderwood.org> wrote:
> > >> You can use a separate field for title aliases. That is what I did for
> > Netflix search.
> > >>
> > >> Why disable idf? Disabling tf for titles can be a good idea, for
> > example the movie “New York, New York” is not twice as much about New
> York
> > as some other film that just lists it once.
> > >>
> > >> Also, consider using a popularity score as a boost.
> > >>
> > >> wunder
> > >> Walter Underwood
> > >> wunder@wunderwood.org
> > >> http://observer.wunderwood.org/ (my blog)
> > >>
> > >>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
> > >>>
> > >>> Hi,
> > >>> We are using solr for our movie title search.
> > >>>
> > >>>
> > >>> As it is "title search", this should be treated different than the
> > normal
> > >>> document search.
> > >>> Hence, we use a modified version of TFIDFSimilarity with the
> following
> > >>> changes.
> > >>> - disabled TF & IDF and will only have 1 as value.
> > >>> - disabled norms by specifying omitNorms as true for all the fields.
> > >>>
> > >>> There are 6 fields with different analyzers and we make use of
> > different
> > >>> weights in edismax's qf & pf parameters to match tokens & boost
> > phrases.
> > >>>
> > >>> But, movies could have aliases and have multiple titles. So, we made
> > the
> > >>> fields multivalued.
> > >>>
> > >>> Now, consider the following four documents
> > >>> 1> "Beauty and the Beast"
> > >>> 2> "The Real Beauty and the Beast"
> > >>> 3> "Beauty and the Beast", "La bella y la bestia"
> > >>> 4> "Beauty and the Beast"
> > >>>
> > >>> Note: Document 3 has two titles in it.
> > >>>
> > >>> So, for a query "Beauty and the Beast" and with the above
> > configuration all
> > >>> the documents receive same score. But 1,3,4 should have got same
> score
> > and
> > >>> document 2 lesser than others.
> > >>>
> > >>> To solve this, we followed what is suggested in the following thread:
> > >>> http://lucene.472066.n3.nabble.com/Influencing-scores-
> > on-values-in-multiValue-fields-td1791651.html
> > >>>
> > >>> Now, the fields which are used to boost are made to use Norms. And
> for
> > >>> matching norms are disabled. This is to make sure that exact & near
> > exact
> > >>> matches are rewarded.
> > >>>
> > >>> But, for the same query, we get the following results.
> > >>> query: "Beauty & the Beast"
> > >>> Search Results:
> > >>> 1> "Beauty and the Beast"
> > >>> 4> "Beauty and the Beast"
> > >>> 2> "The Real Beauty and the Beast"
> > >>> 3> "Beauty and the Beast", "La bella y la bestia"
> > >>>
> > >>> Clearly, the changes have solved only a part of the problem. The
> > document 3
> > >>> should be ranked/scored higher than document 2.
> > >>>
> > >>> This is because lucene considers the total field length across all
> the
> > >>> values in a multivalued field for normalization.
> > >>>
> > >>> How do we handle this scenario and make sure that in multivalued
> > fields the
> > >>> normalization is taken care of?
> > >>>
> > >>>
> > >>> --
> > >>> Regards,
> > >>> Sravan
> > >>
> >
>
--
Regards,
Sravan
Re: Title Search scoring issues with multivalued field & norm
Posted by Tim Casey <tc...@gmail.com>.
For smaller length documents TFIDFSimilarity will weight towards shorter
documents. Another way to say this, if your documents are 5-10 terms, the
5 terms are going to win.
You might think about having per token, or token pair, weight. I would be
surprised if there was not something similar out there. This is a common
issue with any short text.
I guess I would think of this as TFICF, where the CF is the corpus
frequency. You also might want to weight inversely proportional to the age
of the title, older are less important. This is assuming people are doing
searches within some time cluster, newer is more likely.
For some obvious advice, things you probably already know. This kind of
search needs some hard measurement to begin to know how to tune it. You
need to find a reasonable annotated representation. So, if you took the
previous months searches where there is a chain of successive searches. If
you weighted things differently would you shorten the length of the chain.
Can you get the click throughs to happen sooner.
Anyway, just my 2 cents....
On Wed, Jan 31, 2018 at 6:38 PM, Sravan Kumar <sr...@caavo.com> wrote:
>
> @Walter: We have 6 fields declared in schema.xml for title each with
> different type of analyzer. One without processing symbols, other stemmed
> and other removing symbols, etc. So, if we have separate fields for each
> alias it will be that many times the number of final fields declared in
> schema.xml. And we exactly do not know what is the maximum number of
> aliases a movie can have.
> @Walter: I will try this but isn’t there any other way where I can tweak ?
>
> @eric: will try this. But it will work only for exact matches.
>
>
> > On Jan 31, 2018, at 10:39 PM, Erick Erickson <er...@gmail.com>
> wrote:
> >
> > Or use a boost for the phrase, something like
> > "beauty and the beast"^5
> >
> >> On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <
> wunder@wunderwood.org> wrote:
> >> You can use a separate field for title aliases. That is what I did for
> Netflix search.
> >>
> >> Why disable idf? Disabling tf for titles can be a good idea, for
> example the movie “New York, New York” is not twice as much about New York
> as some other film that just lists it once.
> >>
> >> Also, consider using a popularity score as a boost.
> >>
> >> wunder
> >> Walter Underwood
> >> wunder@wunderwood.org
> >> http://observer.wunderwood.org/ (my blog)
> >>
> >>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
> >>>
> >>> Hi,
> >>> We are using solr for our movie title search.
> >>>
> >>>
> >>> As it is "title search", this should be treated different than the
> normal
> >>> document search.
> >>> Hence, we use a modified version of TFIDFSimilarity with the following
> >>> changes.
> >>> - disabled TF & IDF and will only have 1 as value.
> >>> - disabled norms by specifying omitNorms as true for all the fields.
> >>>
> >>> There are 6 fields with different analyzers and we make use of
> different
> >>> weights in edismax's qf & pf parameters to match tokens & boost
> phrases.
> >>>
> >>> But, movies could have aliases and have multiple titles. So, we made
> the
> >>> fields multivalued.
> >>>
> >>> Now, consider the following four documents
> >>> 1> "Beauty and the Beast"
> >>> 2> "The Real Beauty and the Beast"
> >>> 3> "Beauty and the Beast", "La bella y la bestia"
> >>> 4> "Beauty and the Beast"
> >>>
> >>> Note: Document 3 has two titles in it.
> >>>
> >>> So, for a query "Beauty and the Beast" and with the above
> configuration all
> >>> the documents receive same score. But 1,3,4 should have got same score
> and
> >>> document 2 lesser than others.
> >>>
> >>> To solve this, we followed what is suggested in the following thread:
> >>> http://lucene.472066.n3.nabble.com/Influencing-scores-
> on-values-in-multiValue-fields-td1791651.html
> >>>
> >>> Now, the fields which are used to boost are made to use Norms. And for
> >>> matching norms are disabled. This is to make sure that exact & near
> exact
> >>> matches are rewarded.
> >>>
> >>> But, for the same query, we get the following results.
> >>> query: "Beauty & the Beast"
> >>> Search Results:
> >>> 1> "Beauty and the Beast"
> >>> 4> "Beauty and the Beast"
> >>> 2> "The Real Beauty and the Beast"
> >>> 3> "Beauty and the Beast", "La bella y la bestia"
> >>>
> >>> Clearly, the changes have solved only a part of the problem. The
> document 3
> >>> should be ranked/scored higher than document 2.
> >>>
> >>> This is because lucene considers the total field length across all the
> >>> values in a multivalued field for normalization.
> >>>
> >>> How do we handle this scenario and make sure that in multivalued
> fields the
> >>> normalization is taken care of?
> >>>
> >>>
> >>> --
> >>> Regards,
> >>> Sravan
> >>
>
Re: Title Search scoring issues with multivalued field & norm
Posted by Sravan Kumar <sr...@caavo.com>.
@Walter: We have 6 fields declared in schema.xml for title each with different type of analyzer. One without processing symbols, other stemmed and other removing symbols, etc. So, if we have separate fields for each alias it will be that many times the number of final fields declared in schema.xml. And we exactly do not know what is the maximum number of aliases a movie can have.
@Walter: I will try this but isn’t there any other way where I can tweak ?
@eric: will try this. But it will work only for exact matches.
> On Jan 31, 2018, at 10:39 PM, Erick Erickson <er...@gmail.com> wrote:
>
> Or use a boost for the phrase, something like
> "beauty and the beast"^5
>
>> On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <wu...@wunderwood.org> wrote:
>> You can use a separate field for title aliases. That is what I did for Netflix search.
>>
>> Why disable idf? Disabling tf for titles can be a good idea, for example the movie “New York, New York” is not twice as much about New York as some other film that just lists it once.
>>
>> Also, consider using a popularity score as a boost.
>>
>> wunder
>> Walter Underwood
>> wunder@wunderwood.org
>> http://observer.wunderwood.org/ (my blog)
>>
>>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
>>>
>>> Hi,
>>> We are using solr for our movie title search.
>>>
>>>
>>> As it is "title search", this should be treated different than the normal
>>> document search.
>>> Hence, we use a modified version of TFIDFSimilarity with the following
>>> changes.
>>> - disabled TF & IDF and will only have 1 as value.
>>> - disabled norms by specifying omitNorms as true for all the fields.
>>>
>>> There are 6 fields with different analyzers and we make use of different
>>> weights in edismax's qf & pf parameters to match tokens & boost phrases.
>>>
>>> But, movies could have aliases and have multiple titles. So, we made the
>>> fields multivalued.
>>>
>>> Now, consider the following four documents
>>> 1> "Beauty and the Beast"
>>> 2> "The Real Beauty and the Beast"
>>> 3> "Beauty and the Beast", "La bella y la bestia"
>>> 4> "Beauty and the Beast"
>>>
>>> Note: Document 3 has two titles in it.
>>>
>>> So, for a query "Beauty and the Beast" and with the above configuration all
>>> the documents receive same score. But 1,3,4 should have got same score and
>>> document 2 lesser than others.
>>>
>>> To solve this, we followed what is suggested in the following thread:
>>> http://lucene.472066.n3.nabble.com/Influencing-scores-on-values-in-multiValue-fields-td1791651.html
>>>
>>> Now, the fields which are used to boost are made to use Norms. And for
>>> matching norms are disabled. This is to make sure that exact & near exact
>>> matches are rewarded.
>>>
>>> But, for the same query, we get the following results.
>>> query: "Beauty & the Beast"
>>> Search Results:
>>> 1> "Beauty and the Beast"
>>> 4> "Beauty and the Beast"
>>> 2> "The Real Beauty and the Beast"
>>> 3> "Beauty and the Beast", "La bella y la bestia"
>>>
>>> Clearly, the changes have solved only a part of the problem. The document 3
>>> should be ranked/scored higher than document 2.
>>>
>>> This is because lucene considers the total field length across all the
>>> values in a multivalued field for normalization.
>>>
>>> How do we handle this scenario and make sure that in multivalued fields the
>>> normalization is taken care of?
>>>
>>>
>>> --
>>> Regards,
>>> Sravan
>>
Re: Title Search scoring issues with multivalued field & norm
Posted by Erick Erickson <er...@gmail.com>.
Or use a boost for the phrase, something like
"beauty and the beast"^5
On Wed, Jan 31, 2018 at 8:43 AM, Walter Underwood <wu...@wunderwood.org> wrote:
> You can use a separate field for title aliases. That is what I did for Netflix search.
>
> Why disable idf? Disabling tf for titles can be a good idea, for example the movie “New York, New York” is not twice as much about New York as some other film that just lists it once.
>
> Also, consider using a popularity score as a boost.
>
> wunder
> Walter Underwood
> wunder@wunderwood.org
> http://observer.wunderwood.org/ (my blog)
>
>> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
>>
>> Hi,
>> We are using solr for our movie title search.
>>
>>
>> As it is "title search", this should be treated different than the normal
>> document search.
>> Hence, we use a modified version of TFIDFSimilarity with the following
>> changes.
>> - disabled TF & IDF and will only have 1 as value.
>> - disabled norms by specifying omitNorms as true for all the fields.
>>
>> There are 6 fields with different analyzers and we make use of different
>> weights in edismax's qf & pf parameters to match tokens & boost phrases.
>>
>> But, movies could have aliases and have multiple titles. So, we made the
>> fields multivalued.
>>
>> Now, consider the following four documents
>> 1> "Beauty and the Beast"
>> 2> "The Real Beauty and the Beast"
>> 3> "Beauty and the Beast", "La bella y la bestia"
>> 4> "Beauty and the Beast"
>>
>> Note: Document 3 has two titles in it.
>>
>> So, for a query "Beauty and the Beast" and with the above configuration all
>> the documents receive same score. But 1,3,4 should have got same score and
>> document 2 lesser than others.
>>
>> To solve this, we followed what is suggested in the following thread:
>> http://lucene.472066.n3.nabble.com/Influencing-scores-on-values-in-multiValue-fields-td1791651.html
>>
>> Now, the fields which are used to boost are made to use Norms. And for
>> matching norms are disabled. This is to make sure that exact & near exact
>> matches are rewarded.
>>
>> But, for the same query, we get the following results.
>> query: "Beauty & the Beast"
>> Search Results:
>> 1> "Beauty and the Beast"
>> 4> "Beauty and the Beast"
>> 2> "The Real Beauty and the Beast"
>> 3> "Beauty and the Beast", "La bella y la bestia"
>>
>> Clearly, the changes have solved only a part of the problem. The document 3
>> should be ranked/scored higher than document 2.
>>
>> This is because lucene considers the total field length across all the
>> values in a multivalued field for normalization.
>>
>> How do we handle this scenario and make sure that in multivalued fields the
>> normalization is taken care of?
>>
>>
>> --
>> Regards,
>> Sravan
>
Re: Title Search scoring issues with multivalued field & norm
Posted by Walter Underwood <wu...@wunderwood.org>.
You can use a separate field for title aliases. That is what I did for Netflix search.
Why disable idf? Disabling tf for titles can be a good idea, for example the movie “New York, New York” is not twice as much about New York as some other film that just lists it once.
Also, consider using a popularity score as a boost.
wunder
Walter Underwood
wunder@wunderwood.org
http://observer.wunderwood.org/ (my blog)
> On Jan 31, 2018, at 4:38 AM, Sravan Kumar <sr...@caavo.com> wrote:
>
> Hi,
> We are using solr for our movie title search.
>
>
> As it is "title search", this should be treated different than the normal
> document search.
> Hence, we use a modified version of TFIDFSimilarity with the following
> changes.
> - disabled TF & IDF and will only have 1 as value.
> - disabled norms by specifying omitNorms as true for all the fields.
>
> There are 6 fields with different analyzers and we make use of different
> weights in edismax's qf & pf parameters to match tokens & boost phrases.
>
> But, movies could have aliases and have multiple titles. So, we made the
> fields multivalued.
>
> Now, consider the following four documents
> 1> "Beauty and the Beast"
> 2> "The Real Beauty and the Beast"
> 3> "Beauty and the Beast", "La bella y la bestia"
> 4> "Beauty and the Beast"
>
> Note: Document 3 has two titles in it.
>
> So, for a query "Beauty and the Beast" and with the above configuration all
> the documents receive same score. But 1,3,4 should have got same score and
> document 2 lesser than others.
>
> To solve this, we followed what is suggested in the following thread:
> http://lucene.472066.n3.nabble.com/Influencing-scores-on-values-in-multiValue-fields-td1791651.html
>
> Now, the fields which are used to boost are made to use Norms. And for
> matching norms are disabled. This is to make sure that exact & near exact
> matches are rewarded.
>
> But, for the same query, we get the following results.
> query: "Beauty & the Beast"
> Search Results:
> 1> "Beauty and the Beast"
> 4> "Beauty and the Beast"
> 2> "The Real Beauty and the Beast"
> 3> "Beauty and the Beast", "La bella y la bestia"
>
> Clearly, the changes have solved only a part of the problem. The document 3
> should be ranked/scored higher than document 2.
>
> This is because lucene considers the total field length across all the
> values in a multivalued field for normalization.
>
> How do we handle this scenario and make sure that in multivalued fields the
> normalization is taken care of?
>
>
> --
> Regards,
> Sravan