Suppose I use a custom design to create an experiment and set the response goal to “maximize,” with only an upper limit specified as 7. Based on this, an experimental data table is generated, and I fill in response values, all of which are greater than 7. After fitting a model, in the Prediction Pro

[#1]

Suppose I use a custom design to create an experiment and set the response goal to “maximize,” with only an upper limit specified as 7. Based on this, an experimental data table is generated, and I fill in response values, all of which are greater than 7. After fitting a model, in the Prediction Profiler, the desirability settings for the response show default values: a high value of 7, a middle value of 6.5224, and a low value of 6.0448.

My question is:
How are the middle value (6.5224) and the low value (6.0448) calculated? Can they be derived explicitly?

[#2]

Hi @CompositeCamel5,

I don't know the exact details on how the medium and low values are calculated for the desirability function in prediction profiler, but I guess they are calculated by default by considering the range of the response data and its variability. the default values of the desirability profile have little interest, since they do not consider your domain expertise, objective and possible constraints (+ tradeoffs if you have several responses to optimize and they don't converge towards the same optimum). Since you have set an upper limit of 7 with the goal "Maximize", the high value in desirability is set automatically at 7.

You can define and set the desirability profile, values and objective in the red triangle next to the Prediction Profiler. You can set the desirabilities and save them so that they are saved in the datatable in Response Limits column property of your response.

Hope this answer will help you,