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 "venkatesham.gundu@igate.com" <ve...@igate.com> on 2013/06/14 11:49:54 UTC
How spell checker used if indexed document is containing misspelled
words
My data is picked from social media sites and misspelled words are very
frequent in social text because of the informal mode of
communication.Spellchecker does not work here because misspelled words are
present in the text corpus and not in the search query. Finding documents
with all the different misspelled forms for a given word is not possible
using spellchecker, how to go ahead with this.
--
View this message in context: http://lucene.472066.n3.nabble.com/How-spell-checker-used-if-indexed-document-is-containing-misspelled-words-tp4070463.html
Sent from the Solr - User mailing list archive at Nabble.com.
Re: How spell checker used if indexed document is containing misspelled
words
Posted by Upayavira <uv...@odoko.co.uk>.
You're wanting to make your search more fuzzy. You could try phonetic
search, but that's very fuzzy. Go to the analysis tab in the admin UI.
Locate the 'phonetic' field type in the drop down, and you can see what
will happen to terms when they are converted to phonetic equivalents.
Upayavira
On Fri, Jun 28, 2013, at 10:29 AM, venkatesham.gundu@igate.com wrote:
> Thanks for the replies.
>
> I have already tried options mentioned here, apparently those provide
> suggestions for the query word which is incorrectly spelled. I am looking
> a
> feature that - my query term is correct and I want the results in those
> documents both correct spelled term matches and incorrect spelled term
> matches if any -
>
> For example I am searching for documents which have the term "headache",
> now
> I want search results all the documents which have been correct spelled
> headache and also any documents which have been a wrong spelling or typos
> of
> headache like "headcahe" and so on.
>
> Please provide any thoughts and ideas to resolve this problem.
>
> Thanks.
>
>
>
> --
> View this message in context:
> http://lucene.472066.n3.nabble.com/How-spell-checker-used-if-indexed-document-is-containing-misspelled-words-tp4070463p4073879.html
> Sent from the Solr - User mailing list archive at Nabble.com.
Re: How spell checker used if indexed document is containing
misspelled words
Posted by "venkatesham.gundu@igate.com" <ve...@igate.com>.
Thanks for the replies.
I have already tried options mentioned here, apparently those provide
suggestions for the query word which is incorrectly spelled. I am looking a
feature that - my query term is correct and I want the results in those
documents both correct spelled term matches and incorrect spelled term
matches if any -
For example I am searching for documents which have the term "headache", now
I want search results all the documents which have been correct spelled
headache and also any documents which have been a wrong spelling or typos of
headache like "headcahe" and so on.
Please provide any thoughts and ideas to resolve this problem.
Thanks.
--
View this message in context: http://lucene.472066.n3.nabble.com/How-spell-checker-used-if-indexed-document-is-containing-misspelled-words-tp4070463p4073879.html
Sent from the Solr - User mailing list archive at Nabble.com.
RE: How spell checker used if indexed document is containing
misspelled words
Posted by "Dyer, James" <Ja...@ingramcontent.com>.
There are two newer parameters that work better than "onlyMorePopular":
spellcheck.alternativeTermCount
- This is the # of suggestions you want for terms that exist in the index. You can set it the same as "spellcheck.count", or less if you don't want as many suggestions for these.
http://wiki.apache.org/solr/SpellCheckComponent#spellcheck.alternativeTermCount
spellcheck.maxResultsForSuggest
- This lets you give a "did-you-mean" suggestion if the query only gets a few hits. Useful if the user enters a misspelled terms that is in the index but could have gotten a lot more results if they had spelled it correctly.
http://wiki.apache.org/solr/SpellCheckComponent#spellcheck.maxResultsForSuggest
James Dyer
Ingram Content Group
(615) 213-4311
-----Original Message-----
From: Shreejay [mailto:shreejayn@gmail.com]
Sent: Friday, June 14, 2013 8:38 AM
To: solr-user@lucene.apache.org
Subject: Re: How spell checker used if indexed document is containing misspelled words
Hi,
Have you tried this?
http://wiki.apache.org/solr/SpellCheckComponent#spellcheck.onlyMorePopular
Of course this is assuming that your corpus has correct words occurring more frequently than incorrect ones!
--
Shreejay
On Friday, June 14, 2013 at 2:49, venkatesham.gundu@igate.com wrote:
> My data is picked from social media sites and misspelled words are very
> frequent in social text because of the informal mode of
> communication.Spellchecker does not work here because misspelled words are
> present in the text corpus and not in the search query. Finding documents
> with all the different misspelled forms for a given word is not possible
> using spellchecker, how to go ahead with this.
>
>
>
> --
> View this message in context: http://lucene.472066.n3.nabble.com/How-spell-checker-used-if-indexed-document-is-containing-misspelled-words-tp4070463.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
>
Re: How spell checker used if indexed document is containing
misspelled words
Posted by Shreejay <sh...@gmail.com>.
Hi,
Have you tried this?
http://wiki.apache.org/solr/SpellCheckComponent#spellcheck.onlyMorePopular
Of course this is assuming that your corpus has correct words occurring more frequently than incorrect ones!
--
Shreejay
On Friday, June 14, 2013 at 2:49, venkatesham.gundu@igate.com wrote:
> My data is picked from social media sites and misspelled words are very
> frequent in social text because of the informal mode of
> communication.Spellchecker does not work here because misspelled words are
> present in the text corpus and not in the search query. Finding documents
> with all the different misspelled forms for a given word is not possible
> using spellchecker, how to go ahead with this.
>
>
>
> --
> View this message in context: http://lucene.472066.n3.nabble.com/How-spell-checker-used-if-indexed-document-is-containing-misspelled-words-tp4070463.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
>