Hi, I am a developer with experience in Natural Language Processing.I have conceptualized a system for Libreoffice Writer and would like to implement it.
The system I have conceptualized is: The system shall be a semantic checker for Writer.Spell checking systems have the facility of flagging erratical spellings and suggest useful correctiions in most cases.But,we do not have a system which can perform semantic checking and corrections.For e.g. if a user types 'I apples like' instead of 'I like apples',the spell check will not find any error in it,but,our new system will be able to flag the error and suggest corrections. This shall be done using fuzzy string matching and a knowledge base of sentence models.The sentence models shall be of format <noun><identifier><adjective> etc. The system shall be in a learning mode initially.Users can point out mistakes in the error detection done by the system and the system will make a not of the correction by the user. Eventually,with experience,the system will accurately point out grammatical errors and try to suggest corrections for them. I have designed a system in the past which evaluates students subjective answers to questions based on a model answer given by the faculty and then,rates the answer. I really want to implement this system for Libreoffice.Please let me know if I can proceed further. Thanks a lot in advance, Atri _______________________________________________ LibreOffice mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/libreoffice
