nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Neural Networks - modeling details

marta-simoes

Dear all,

I was getting deeper into the neural modeling in JMP, but I don't find any info on:

- training times?

- loss in function of iteration/epoch for the model fitting?

Any help?

Happy Easter to everyone,

Kind regards,

Marta

Accepted answers

Mark_Bailey

See the attached JMP white paper for methodological details.

JMP Neural Network Methodology.pdf

All comments

Mark_Bailey

See the attached JMP white paper for methodological details.

JMP Neural Network Methodology.pdf

marta-simoes

Thanks Mark!

MathStatChem

You can also check out the book Building Better Models with JMP Pro which is currently available for free (see https://support.sas.com/en/books.html) from SAS Press. Chapter 7 is on Neural Networks.

Ese_Omatsone

I don't have JMP Pro. I have JMP 16.

My question is - how/where can I change the output parameters that JMP provides after completing a neural network analysis. I''m especially interested in seeing the RMSE and MAPE. Or are RMSE and RASE the same thing?

Mark_Bailey

The standard report provides the RMSE. See the documentation. You do not need JMP Pro for this result.

You will not be able to use a data column in the validation role, use the informative missing option, more than the hyperbolic tangent activation function, or more than one hidden layer without JMP Pro.