please help with dim, guessing chinese input.
Hi,
dim.sourceforge.net debian chinese input method.
the dim development has being stalled for awhile, partially because
i'm busying doing cdlinux stuff, partially because i have some
technically difficulties with dim guessing functions. pleaze help!
thanks! :)
my origal idea to guess from a pinyin string to a chinese char string
is similiar to what i 've got for learnning functions. which i'm
pretty satisfied right now ('till some statistcs fuck me up)
the basic ideas behind learning func is that:
1. most chinese words are two char words
2. chinese sentence structure is much more inportant than a link of
words.
the second idea make me decided to drop a markov link alike algo
sugessted by my brother earlier.
with these ideas behind, the learning functions could learn from user
input or chinese articales, many (two char or more) words and many
sentence structures which is basically ...XXX___XXX___XXX... that is
some blanks to be filled by some other regognized chinese words.
but i feel to derive guessing functions, these ideas are not enough. i
nearly go to buy some chinese grammar book, hehe.. (later maybe) i
hope some of you gurus here could shed some light on me.
i have a rough idea that i sould do a modularized guessing, but
haven't got time enough to think it through.
please help! thanks!
p.s. statistics on chinese grammer, urls etc. ideas, i don't need
bullet proof ideas. ;) program deisg to let further
evolve. etc. please! thanks!
--
http://dim.sourceforge.net ............... Debian Chinese Input Method
http://njlug.sourceforge.net ............ NanJing GNU/Linux User Group
http://cdlinux.sourceforge.net ........... Debian running on Live! CDs
http://people.debian.org/~zw ...................... XEmacs Screenshots
Reply to: