JMP Academic Series: Teaching Design of Experiments using JMP (23 Feb 2017)
In this webinar we demonstrate tools in JMP to make teaching the design of experiments most effective. We show classical and custom designs as well as modern and spezial purpose designs. Tips and resources are shared for teaching DOE. The JMP journal with all links can also be download in the JMP Academic Webinar Library at www.JMP.com/teach.
Direct links to sections:
Get Started DOE • JMP Academic Series: T...
Classical Designs • JMP Academic Series: T...
Custom Designs • JMP Academic Series: T...
Advanced DOE • JMP Academic Series: T...
Пікірлер: 6
very good. helped me so much. in "design generation", 9:06 min, what means "group runs into ramdom blocks of size"?
Dear Dr Kraft. Greetings from Germany. When I calculate DOE in JMP I am often facing the Problem that I get (for example for a factor) a value with a lot of digits after the comma. How can I avoid that? I can not adjust these values in that accuracy in my test. When I did the design can I overwrite the values with the next possibles? (This happens when I do RSM Designs, were JMP calculates intermediate steps) Thanks for the answers.
@volkerkraft1984
4 жыл бұрын
Hi, you can change the displayed format and digits in the column settings (double-click the column header). Or you can change the values by removing digits or rounding values. In case you are concerned about how these changes effect your design, you can compare the design diagnostics between the new and your original design.
@TheWechris
4 жыл бұрын
Volker Kraft Hi thanks, for the answer! Do you also have a recommendation for a similar problem: When we do a DOE we have to manufacture special parts for testing (we are in fine Mechanics Maschine building) for each test setting we need about 100 pcs. So we try to hit the calculated values in the DOE (mean value of the test sample) but that is not guaranteed because of the normal manufacturing deviation. So I would measure parts of every Plot (Split Plot Design / 4 hard to change factors) and replace the measured mean value with the calculated values in the DOE. Is that a good way for getting the desired output (the Modell) or do you recommend a different approach? Thanks again.
@volkerkraft1984
4 жыл бұрын
TheWechris If it is hard or even impossible to find samples with the right spec required by your design, you can consider covariate factors in Custom Design. JMP can also select the best subset if you have samples to choose from. More at JMP.com/help or email volker.kraft@JMP.com .
@TheWechris
4 жыл бұрын
Volker Kraft thanks again. Seems I need more training in this matters with JMP. Now I have some more Ideas which could be done to solve the problem. Have a good day and stay save! 👍