You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The logic of checking "has_directed_path", as used in AutoIdentifier.py::identify_effect() is that it is an early sanity check that the caller is asking a sensible question. I.e., if a variable X has no way of influencing Y according to the causal graph, then it doesn't make sense to do a fancy calculation to give you the same response. It's perhaps more likely that the caller is making a mistake somewhere and we want to call that out.
it only supports singleton treatment and outcome sets, but might be called with sets that have greater cardinality
So, if we have a higher cardinality, I can see two options:
(1) a minimum requirement would be that at least one of the treatments has a direct path to at least one of the outcomes.
(2) more aggressively: every one of the treatments has at least one direct path to one of the outcomes. And, every one of the outcomes has a direct path from at least one of the treatments.
The logic of checking "has_directed_path", as used in AutoIdentifier.py::identify_effect() is that it is an early sanity check that the caller is asking a sensible question. I.e., if a variable X has no way of influencing Y according to the causal graph, then it doesn't make sense to do a fancy calculation to give you the same response. It's perhaps more likely that the caller is making a mistake somewhere and we want to call that out.
it only supports singleton treatment and outcome sets, but might be called with sets that have greater cardinality
So, if we have a higher cardinality, I can see two options:
(1) a minimum requirement would be that at least one of the treatments has a direct path to at least one of the outcomes.
(2) more aggressively: every one of the treatments has at least one direct path to one of the outcomes. And, every one of the outcomes has a direct path from at least one of the treatments.
I'd probably go for #2
Originally posted by @emrekiciman in #640 (comment)
The text was updated successfully, but these errors were encountered: