Implement Javascript functions turn rules into unknown word recognizer (Inuktitut Spell Checker Iteration 3 - Adding rules to "recognize" more words)

[GoogleCodeExporter]

Purpose of implementation request:
To try something very challenging at the interface of X-bar theory and Finite 
State Machines.

When implementing the request, please focus on these
steps/functions/components:
  * design a module (it will be rather complicated) to break unrecognized words into morphemes, and ask the user if that is why the word wasn't recognized (ex: the word tokenization is not in most English wordlists, but it is an English word. In this module you would present token-ize-ation to the user and they could approve the word and add it to the generative lexicon) (3 hours)
  * get feedback on your design (2 hours)
  * try to implement your design (10 hours)
  * try to find existing projects which implement parts of your design (4 hours) [NRC's inuktitutcomputing.ca, hunspell, SIL's ToolBox, Goldsmith's Linguistica, Creutz' Morphessor]
  * try to compile/run those existing projects (15 hours)
  * pair program your module using whatever you have learned (10 hours)

When implementing the request, watch out for the following potential
security/lack of access/lack of data/formatting etc hiccups:
  * this will be a near impossible task (it requires expertise in NLP, Machine Learning and Software Engineering)

Expected next steps see Iteration 4 - Re-factor word recognition to use the API 
for the NRC's morphological analyser

Original issue reported on code.google.com by [email protected] on 25 Nov 2011 at 7:10

• Blocked on: Implement Javascript functions to "learn" new words (Inuktitut Spell Checker Iteration 2 - Making the script adaptive) #24

[GoogleCodeExporter]

Original comment by [email protected] on 25 Nov 2011 at 7:12

• Added labels: Type-Implementation