Is your feature request related to a problem?
For sepsis models, its often desirable to do lead-time analysis by looking at the first time a model score exceeds a given threshold.
Right now "first" aggregates the first score in the window, and "max" takes the largest value within the window.
Describe the solution you'd like
If we can do "first_above_threshold" with a given threshold, that would be helpful.
Additional context
One downside to this aggregation is that it requires a specific threshold, so some of the aggregation may need to pass that value through (current controls allow more than one threshold to see how well a model works at different thresholds, which might cause confusion as to which threshold relates)
Is your feature request related to a problem?
For sepsis models, its often desirable to do lead-time analysis by looking at the first time a model score exceeds a given threshold.
Right now "first" aggregates the first score in the window, and "max" takes the largest value within the window.
Describe the solution you'd like
If we can do "first_above_threshold" with a given threshold, that would be helpful.
Additional context
One downside to this aggregation is that it requires a specific threshold, so some of the aggregation may need to pass that value through (current controls allow more than one threshold to see how well a model works at different thresholds, which might cause confusion as to which threshold relates)