Why JMP Pro ?Hi all,
We got urgent suggestion 'Why JM Pro' from big customer.
They are Display manufacturer that produce LCD, OLED etc.
They have used JMP for several years and they want to pursuade their manage
and director to buy JMP Pro.
JMP Pro pamphlet isn't sufficient to them
If you have some material, report and/or success case to JMP Pro,
please let me share.
Thanks
Ikju
Re: Why JMP Pro ?Yes, I agree with @martindemel. FDE is the strongest reason for a customer to go with JMP Pro.
I understand that LCD manufacturers use dispersions of pigment. So they might be able to relate to the challenge of pigment milling processes as examples. As Martin says, I have created two demos for JMP Pro that use pigment milling as examples. This is one of them using FDE:
https://community.jmp.com/t5/Phil-Kay-s-Blog/Learn-More-from-Your-Experiments-with-Functional-Data-Explorer-3/ba-p/82777
You could ask them if they have functional data (e.g. sensors measuring as a function of time). And if they need to understand how this functional data relates to process inputs ("Functional DOE", example above) or process outputs ("Functional Machine Learning") .
This is an example of Functional Machine Learning:
https://community.jmp.com/t5/Phil-Kay-s-Blog/Understanding-Complex-Yield-Issues-with-Sensor-Data-5-Minute/ba-p/84620
I hope this helps.
Phil
Re: Why JMP Pro ?Hi Ikju,
One topic might be functional DOE. Phil's example of a milling process is (I believe) based on a milling process for a LCD Display. It is about reduciing the time (and if you want energy) spend to mill down the particles below a threshold in lower hundreds nanometer.
You can be satisfied with the point of time to get into spec. Or you want to study the profile to get into spec and stay there for long which makes the process way more stable. If they do as they did before they probably ignore all the data and therefore information collected over the time the process runs. May they also have an ideal process they would like to see the milling happen, then they could use the golden curve analysis to find a setting closest to their ideal curve. All this is about Functional Data Explorer and Generalized Regression.
In the milling example he also points out that with strong correlations between factors and responses you might not be able to identify the true factors to investigate on further (e.g. if the temperature has a strong relationship with the time needed to mill down the particles it might hide other important factors like pigment% or beeds% though technically very important, just because of multicolinearity effects) using standard fitting methods.You can work arround this e.g. using advanced methods like bootstrap forest to find more potential important predictors to be used in a DOE (e.g. DSD or Custom Design) to find the true underlying relationships.
@Phil_Kay might can come up with additional resasons for JMP Pro.
Another topic might be (but this is at the beginning) using SEM to better understand the underlying relationships in a manufacturing process.
