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
Answering @scranefield question on how to test stuff, we need to decide:
Felipe which is the best experiment to run to evaluate the effects of learning p_non_comp on norm learning?
My hypothesis is that if we get p_non_comp wrong, it depends on the value of p_sanc, I can think of six cases:
p_sanc high, p_non_comp correct: ideal case, we should get high high-precision, high-recall
p_sanc low, p_non_comp correct: we may overestimate some norms, so recall should be OK
p_sanc high, p_non_comp low: we may get low recall (because we won't see much non-compliant behaviour)
p_sanc low, p_non_comp low: we may get low recall (because we won't see much non-compliant behaviour), and low precision because the ones we do get may not get sanctioned
p_sanc high, p_non_comp high: we should get low precision because we will mistakenly think a lot of the generated behaviour is non-compliant, even when we don't see sanctioning
p_sanc low, p_non_comp high: we may get really high low recall (because we will see non-compliant behaviour), but low precision because we will mistakenly think a lot of the correct behaviour is non-compliant.
The text was updated successfully, but these errors were encountered:
Answering @scranefield question on how to test stuff, we need to decide:
My hypothesis is that if we get
p_non_comp
wrong, it depends on the value ofp_sanc
, I can think of six cases:p_sanc
high,p_non_comp
correct: ideal case, we should get high high-precision, high-recallp_sanc
low,p_non_comp
correct: we may overestimate some norms, so recall should be OKp_sanc
high,p_non_comp
low: we may get low recall (because we won't see much non-compliant behaviour)p_sanc
low,p_non_comp
low: we may get low recall (because we won't see much non-compliant behaviour), and low precision because the ones we do get may not get sanctionedp_sanc
high,p_non_comp
high: we should get low precision because we will mistakenly think a lot of the generated behaviour is non-compliant, even when we don't see sanctioningp_sanc
low,p_non_comp
high: we may get really high low recall (because we will see non-compliant behaviour), but low precision because we will mistakenly think a lot of the correct behaviour is non-compliant.The text was updated successfully, but these errors were encountered: