commonsense team mailing list archive

[Ken Arnold <kenneth.arnold@xxxxxxxxx>]

Public bug reported:

We have a number of passive concepts, like "get fired" and "getting
served". They should be treated differently than their active
counterparts, e.g., "fire" and "serve". "getting served" normalizes to
"get serve", which seems right, but "get fired" normalizes to "fire".

I suspect "get" is a stopword but "getting" isn't, and stopword removal
happens before lemmatization. Fixing this bug will require a test that
clearly illustrates the desired behavior lest it break again ;)

** Affects: conceptnet
     Importance: Undecided
         Status: New


** Tags: parsing

-- 
get/getting +ed
https://bugs.launchpad.net/bugs/469606
You received this bug notification because you are a member of
Commonsense Computing, which is the registrant for ConceptNet.

Status in ConceptNet API: New

Bug description:
We have a number of passive concepts, like "get fired" and "getting served". They should be treated differently than their active counterparts, e.g., "fire" and "serve". "getting served" normalizes to "get serve", which seems right, but "get fired" normalizes to "fire".

I suspect "get" is a stopword but "getting" isn't, and stopword removal happens before lemmatization. Fixing this bug will require a test that clearly illustrates the desired behavior lest it break again ;)