Closed ncatlett closed 6 years ago
I followed the directions in the tutorial PDF for making the pull request for a new BEP - I think I may have deleted the template.
I like the relationship proposal and like how it handles ambiguous correlations, but I prefer to not have an abbreviation for it - at least 'no' doesn't work for me when reading through a sample relationship.
agree that 'no' does not work as an abbreviation
Also agree that no
is a bad abbreviation. I think cor
is the abbreviation for the non-directional correlation
relationship, so maybe noCor
?
What do you guys think about the idea of "negative" knowledge as a more general concept?
There's already the causesNoChange
relationship, which is effectively the negative of regulates
. noCorrelation
would be the negative of the correlation
relationship.
I think it's the case that there are a lot of times there is a "negative" version of a statement (even direct and causal relations) in the literature, but we don't curate it because it's just not that useful. BEL isn't always oriented towards utility, but rather just encoding what's possible. Conversely, I would imagine representing negative knowledge in a network would be difficult, and so would reasoning over it.
Approved in a vote by the BEP Review team - John Bachman, Charlie Hoyt and William Hayes.
add noCorrelation relationship