FAQ
Hi.

I'd like to register:

Name: Sergei Golubchik
Email: sergii@pisem.net
Homepage: none at the moment
preferred user-ID: SERG (looks like it's still vacant)

I have Lingua::RU::Translit module that converts from russian
"translit" encoding - russian text written using latin1 charset -
to russian in koi8-r: sub translit2koi

Convertion russian-to-translit is lossy, so this module uses Hidden
Markov Model to find the most probable original russian text.
Leaves Russian in cyrillic or native English as is.

Can be very useful in email applications (email in "translit" Russian
are way too common, because not everybody can or know how setup his mail
client for Cyrillic).
I personally use it in xchat plugin.

possible todo:

add koi2translit - it's trivial
tr/.../.../; and few s/.../.../g; everybody can write it in a minute,
but I could add it here for completeness :)

Extend decoder by
* addind feedback
* loadable transition tables
* loadable characted equivalent tables
* more bullet-proof HMM
* spell-checker (or simpe list of exceptions)
* neural net instead of HMM ?

Regards,
Sergei

Search Discussions

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupmodules @
categoriesperl
postedOct 19, '02 at 11:21p
activeOct 19, '02 at 11:21p
posts1
users1
websitecpan.org...

1 user in discussion

Sergei Golubchik: 1 post

People

Translate

site design / logo © 2021 Grokbase