0
$\begingroup$

The question is: can exist a parallel or serial neural network or someone talked about this? For explanation, in the network a single record in a data set enter in the input layer as a value between 0 and 1, but, could i interpretate the value parallely as an array of bits (so i need an input layer as large as the number of bits composing the value) or even serially? or this is a bad technic?

Imagine a number (es. 5) that is interpretated as an n-bits values (101) in a 3-way input layer instead of a degree (0,?) in a 1-way input layer.

I'm a novice in neural network and the question could be stupid. Sorry if it is.

$\endgroup$
0
$\begingroup$

Yes, of course you could do that. Whether it is useful will depend on the task. I'd suggest spending some quality time on tutorials/classes on neural networks, and then spend some time experimenting with different network architectures to see how the choice of feature vector affects the performance of the network.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ As you suggested, i must spend very time to have a good knowledge and it's what i'm trying to do now, but for this doubt i wasn't able to found any good direct answer. Many thanks! If you have some good links, i would appreciate it :D $\endgroup$ – Jump46 Aug 6 '16 at 19:40

Not the answer you're looking for? Browse other questions tagged or ask your own question.