I use prescriptive scoring function to find optimized lever values with my model.
But I am wondering how to use presciptive scoring with some constraints. Some attributes(like quality) must stay in a properly range when min/max the goal. How can I do such works using prescriptive scoring.
I am investigating your question.
In the meantime could you kindly provide information about your Analytics Server version?
The 7.4 Guide states following:
For features marked as levers, the range of values that can be tested during prescriptive scoring is determined by the range feature at the time when the model is trained. New data can be uploaded after the model is trained, but for prescriptive scoring, the new range will not be considered in the lever metrics.