AI Zone Admin Forum Add your forum

Chatbot Conference in the Netherlands on Tue 24, 2018, with international line-up! Check it out

Partially spelled words
 
 

Is it possible to use a partially spelled word rule in a concept? For example:

concept: ~malade (malade* souffrant* agonisant*)

Thanks

 

 
  [ # 1 ]

No there isn’t, but this is where you canonicals come into play.

Your concept should list the canonical forms of each word, and the dictionary will contain each variant and the canonical for each of them.

 

 
  [ # 2 ]

Thanks, I guess by default the dictionary is in English. How to create my own dictionary containing word variants to be handled by my bot (since I’m with french language )? Also how to tell CS to consider my new dictionary?

 

 
  [ # 3 ]

There is a language= command line parameter to set the language (and %language) to use.
This will cause CS to construct certain folder names with that language name.

I don’t think there is specific documentation on the format of the dictionary files, but the existing text versions should give you enough to go on. The document on foreign languages has a bit more information.

If you decide to do this manually and not use TreeTagger, then you will have to do your own pos tagging and setting of canonicals (see the ^setcanon() function). We use TT, so perhaps others can jump in with their experiences.

 

 
  login or register to react