Prediction profiler Importance rating

[#1]

Hi ,

I am wondering the importance rating in the desirability function that we actually set in the prediction profiler, if a response is more important than my other response do i put a value of 1 or i put a bigger number such 10 or 20?

Rgrds

Irfan

[#2]

According to the online documentation:

Importance

Enter a relative weighting for each of several responses in computing the overall desirability function. The Importance value can be any positive number. When no Importance value is specified, JMP treats all responses in a given analysis as having equal Importance values. If there is only one response, it receives Importance 1.

So you need to decide the relative weighting of the responses you have.

[#3 · Accepted solution]

I just want to add to Ian's explanation that you should not use large numbers like 10 or 20. Such a value will practically eliminate the responses with an Importance value of 1 from the optimization.

From JMP Help:

[#4]

HI mark,

Does it mean i can't use 1,2,3 but instead i have to put a value of 0.5,0.2,0.3 according to the importance of each parameters?Do you have an example to share with me

Rgrds

Irfan

[#5]

You can use what ever values you like so long as you set all of them using the same scale.

If you use 1, 2, 3 then the one with a 2 will be twice as important as the 1 and the 3 is three times as important, likewise 10, 20, 30.

But if you only set one of them and use a 10 then the others will still be at their default value of 1 and you will set that one to be 10 times as important as the others.

[#6]

Thanks