Hi,
For a neural network with architecture 2-7-1-1, is it possible to have a NN output as small as -11.0?
If I want the output of a NN to be from 0 to 1, what parameters should I change?
Thanks
yu
Yes, it’s perfectly possible to have such an output, and there is no way to avoid it.
There is no parameter that allows you to change the output type.
What you can do is to manually pass the output through a sigmoid. The following should give reasonable results:
1/(1+e^(1.25-4.5x)),
where x is the NN output.
Cheers,
Christophe.