Interesting problem.
The following paper presented methods based on knowledge-poor features
to guess grammatical gender in several Indo-European languages:
S. Cucerzan and D. Yarowsky. Minimally Supervised Induction of Grammatical
Gender. Proceedings of HLT/NAACL 2003, Edmonton, 2003.
However, your problem seems harder, in that you are working with names,
not nouns, and I suppose that In Chinese you are not going to get any
morphological cue to gender...
Regards,
Marco
-- Marco Baroni SSLMIT, University of Bologna http://sslmit.unibo.it/~baroni
This archive was generated by hypermail 2b29 : Tue Dec 13 2005 - 15:00:48 MET