cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
mujahida
Level I

Singularity details information from DOE data analysis

Dear JMPers,

 

 I encountered an issue, I got data from DOE, when I analyze these data in Fit Mode, then pump up an information of singularity details, I don't understand what happened to my DOE or data? why and how to solve this problem ?

 

 in order to present my problem, here I attached DOE data and my analysis here.

Thank a lot in advance.

 

2 REPLIES 2
shoffmeister
Level V

Re: Singularity details information from DOE data analysis

The singulartiy details occur when there is some degree of collinearity in your model effects. That means: one model effect is correllated with another model effect or with a combination of multiple other model effects.

 

Especially in a DOE-setting this should not happen most of the times. I have seen that in the presence of missing values or when someone added model effects that were not considered while setting up the doe (maybe a doe for linear effects was planed but you are analyzing it using quadratic effects). This might be a real problem for you analysis!

 

I cannot look into your data right now. But my advice would be to check if you have a lot of missing data. If this is true you might want to try to repeat those experiments. Alternatively remove the model effects that appear in the singularity details from the model and accept that you are not able to learn something about them with the given data. 

 

 

Hope this helps. Maybe some1 else can have a closer look into your data and provide more detailed tipps.

Best,

Sebastian

Phil_Kay
Staff

Re: Singularity details information from DOE data analysis

I always recommend checking the quality of your data with Analyze > Distribution before going into modelling. In this case you would find that one of your factor variables, "pressure difference", looks strange because it has been imported as ordinal, character data. It seems this is because of what look typos in the data table (e.g. a value of -0.2.45378). If you fix these problems in the data you should find that you can fit a model without singularities.