I am very new to Neural Networks so please forgive me if this is a trivial question but my research hasn't produced any meaningful answers. I want to have a neural net that has a binary output (yes or no). However, I want those signals triggered on some sort of a threshold. I am having a bit of a hard time explaining myself so please bear with me.
Example: If the input it 10, I don't (necessarily) want the output to trigger "yes". However, if the input it 4, 5, 7, 10 (sequentially) then I want the output to trigger. So in other words, it has to have a memory of what came before.
I think what I need MIGHT be a Recurrent neural network but I am not really sure because if some threshold isn't met, I don't want my output to fire anything at all (so I guess to be technically correct, for each input, I would have three outputs - yes, no and nothing).
Can anyone point me in the right direction? Thank you.
just use multiple input neurons instead of one? :)
What about a training set like this?: (0=nothing, 1=yes 2=no)
0, 0, 0, 10 Output 0
4, 5, 7, 10 Output 1
Log in to post a comment.