tika-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Robert Muir (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (TIKA-369) Improve accuracy of language detection
Date Fri, 04 Nov 2011 02:35:33 GMT

    [ https://issues.apache.org/jira/browse/TIKA-369?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13143701#comment-13143701
] 

Robert Muir commented on TIKA-369:
----------------------------------

{quote}
Cons: unsurprisingly, has trouble with short text.
{quote}

Not any less trouble than competing libraries:
http://blog.mikemccandless.com/2011/10/accuracy-and-performance-of-googles.html

Its interesting if you read their paper, I think the normalizations etc. they made make total
sense
and I can easily see how that would make a big difference on train vs. test when train is
stuff
like wikipedia (which isn't always totally realistic).

I haven't played with their approach for CJK detection but it makes sense to me, would be
great to
see some evaluation results for that case.

On the other hand I think CLD has nice stuff like segmenting per-script (not ambiguous) first
to
eliminate stupidity when a document has multiple scripts (e.g. cyrillic+latin or arabic+latin)..
it would be great if the cybozu impl integrated this approach as well.

                
> Improve accuracy of language detection
> --------------------------------------
>
>                 Key: TIKA-369
>                 URL: https://issues.apache.org/jira/browse/TIKA-369
>             Project: Tika
>          Issue Type: Improvement
>          Components: languageidentifier
>    Affects Versions: 0.6
>            Reporter: Ken Krugler
>            Assignee: Ken Krugler
>         Attachments: Surprise and Coincidence.pdf, lingdet-mccs.pdf, textcat.pdf
>
>
> Currently the LanguageProfile code uses 3-grams to find the best language profile using
Pearson's chi-square test. This has three issues:
> 1. The results aren't very good for short runs of text. Ted Dunning's paper (attached)
indicates that a log-likelihood ratio (LLR) test works much better, which would then make
language detection faster due to less text needing to be processed.
> 2. The current LanguageIdentifier.isReasonablyCertain() method uses an exact value as
a threshold for certainty. This is very sensitive to the amount of text being processed, and
thus gives false negative results for short runs of text.
> 3. Certainty should also be based on how much better the result is for language X, compared
to the next best language. If two languages both had identical sum-of-squares values, and
this value was below the threshold, then the result is still not very certain.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message