Difference between revisions of "Ideas for Google Summer of Code/Improving support for non-standard text input"
Jump to navigation
Jump to search
Line 14: | Line 14: | ||
==Coding challenge== |
==Coding challenge== |
||
+ | |||
+ | * Make a test corpus of non-standard texts for a particular domain (could be IRC, tweets, forums etc.) |
||
+ | * Translate them with Apertium |
||
+ | * Come up with examples of non-standard features that effect translation quality |
||
+ | * Propose ways in which they might be solved. |
||
==Tasks== |
==Tasks== |
Latest revision as of 12:50, 10 March 2014
Create a module that will standardise non-standard input. For example, slang, abbreviations.
Some examples from English[edit]
- Extra space: "he he" (hehe)
- Spacing and hyphen variation: no-one, noone, no one
- Optional hyphen: re-integrate, reintegrate
- Missing apostrophe: shes thinking about it
- Non-standard capitalisation: im thinking about it
- Abbreviated words: fav,
- Emoticons: :)
Coding challenge[edit]
- Make a test corpus of non-standard texts for a particular domain (could be IRC, tweets, forums etc.)
- Translate them with Apertium
- Come up with examples of non-standard features that effect translation quality
- Propose ways in which they might be solved.
Tasks[edit]
- Do a literature review of papers on normalisation of input.