lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From liat oren <oren.l...@gmail.com>
Subject Re: Scores between words. Boosting?
Date Fri, 27 Mar 2009 10:49:43 GMT
Yes, I have scoring, I will start with the 3 that are the closest.

By the way, I saw that Lucene also has Synonyms. Any reason you suggested
Solr?

Thanks,
Liat
2009/3/24 Grant Ingersoll <gsingers@apache.org>

> Do you have any info that helps you narrow down how many to choose, like
> some type of ranking of the synonyms?  I guess I would start smaller, say
> maybe 3, and then evaluate your results with different numbers.
>
>
> On Mar 22, 2009, at 2:40 PM, liat oren wrote:
>
> Ok, thanks. I will look how to use it.
>>
>> 10 words look too many? How many would you suggest?
>>
>> Thanks again,
>> Liat
>>
>> 2009/3/19 Grant Ingersoll <gsingers@apache.org>
>>
>>
>>> On Mar 19, 2009, at 5:13 AM, liat oren wrote:
>>>
>>> I am looking for a quick solution to expand queries so they will look for
>>>
>>>> synonms as well.
>>>> The same way WordNet is doing - it will looks for other words that mean
>>>> the
>>>> same as written in the query.
>>>>
>>>> So Synonyms and WordNet are better categories to describe what I need.
>>>>
>>>> Any idea?
>>>>
>>>>
>>> You might have a look at Solr's SynonymFilter
>>>
>>>
>>>
>>>> Currently what I do, is expending the query in my own code - for every
>>>> word
>>>> in the query I also add the 10 most close words to that one and parse it
>>>> to
>>>> the query parser.
>>>>
>>>>
>>> Any particular reason for 10?  Seems like for large queries, you could
>>> quickly create a very very large query.
>>>
>>>
>>>
>>>> It takes a bit time and I wondered whether I am missing something and it
>>>> can
>>>> be done in a better way.
>>>>
>>>>
>>> What part is slow?  Can you share more about your current solution
>>> code-wise?  If your synonym list is fairly static, you might consider
>>> adding
>>> synonyms during indexing instead.
>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
>>> For additional commands, e-mail: java-user-help@lucene.apache.org
>>>
>>>
>>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

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