12-17-2017 12:20 AM
I try to use SAS Enterprise Miner to build a neural network model. My dataset is relatively small (54 rows x 42 columns).
SAS EM has Autoneural, neural network, HP neural But I want to control regularization, the number of hidden layers and hidden neurons. Either of three cannot completely meet my demand. Is there any other way to work around it?
I am very new to PROCNNET. Is that correct to use "program editor" in SAS EM to perform PROCNNET?
One more question. How to use LIBNAME libref if I use SAS on demand for academic?
Please, someone, help me with those questions.
Thank you very much.
12-18-2017 02:19 PM
This previous post might help you, it gives examples of how to use PROC HPNEURAL and PROC NEURAL in a SAS programming environment: