HPNeural Final Functional Form Logistic Funciton

quigleym — Sun, 07 Jan 2018 02:42:08 GMT

Hi:

Ran HPNeural using 2 hidden layers with 100 hidden nurons and 50 hidden nurons. The final function form, if I use code file=, is of the form

P_target_good_final1 = 1.0 / (1.0 + EXP(-_T1_0));

which is logistic. Is his true of Proc Neural as well?

Thanks

Re: HPNeural Final Functional Form Logistic Funciton

WendyCzika — Mon, 08 Jan 2018 18:01:05 GMT

Yes, that is what's used in PROC NEURAL as well when you have a binary target.