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
It is useful to relate the policy values in plots to problem specific aggregates, for example, in Figure 5 here we show estimates of intervening on n~2500 units. It's possible to achieve this by just scaling your evaluation scores and costs when calling into maq, but a more direct interface would be nicer.
A natural way to do so would be to delegate this choice to plot, e.g,
plot(qini, scale = 2500, ...)
Would plot the policy values and costs multiplied by 2500.
If you want to read off policy values using average_gain you'd have to do the scaling yourself:
2500 * average_gain(qini, spend = something)
(just tagging @swager in case you have any other opinions on this).
The text was updated successfully, but these errors were encountered:
It is useful to relate the policy values in plots to problem specific aggregates, for example, in Figure 5 here we show estimates of intervening on n~2500 units. It's possible to achieve this by just scaling your evaluation scores and costs when calling into
maq
, but a more direct interface would be nicer.A natural way to do so would be to delegate this choice to plot, e.g,
Would plot the policy values and costs multiplied by 2500.
If you want to read off policy values using
average_gain
you'd have to do the scaling yourself:(just tagging @swager in case you have any other opinions on this).
The text was updated successfully, but these errors were encountered: