Basics: How does a neural network work? (Decision) -


first of all: know there similar questions this. wanna know plain basics. somehow miss important.

let's assume have data (x,y) -> z z can 0 or 1 , x,y in [0,1]. wanna train neural network data , desired output should boundary or line or curved line in x,y space splits zeros ones (e.g. male/female or whatever).

so, wanna have 1 hidden layer. guess somehow understand how feed network:

  1. feed x = (x,y) first layer
  2. do computation weights , bias atc.
  3. compute error or loss function
  4. train network e.g. gradiant descent (i.e. updating weights, etc)
  5. yeah, comes problem :d

what output in end? given data set network tries reproduce z values given x,y, right? so, how fit or decision boundary or whatever? stuff 1 "plots" in end in x,y space?

how generate "new data" network, or not possible?

so, main question is: output? , how handle output? , steps logistic regression plot in end (that can found everywhere in internet, not plot in end :d)

i not it:d sorry^^.

thanks in advance,

marcel

so, after discussion friend (who not around last days) think got it.

after training can use network function f(x,y) make new predictions.

to find decision boundary (the hyperplane in x space) can solve f(x,y) = 1/2 x(y) or vice versa , plot it. done:d

cheers,

marcel speaking myself :d


Comments

Popular posts from this blog

javascript - jQuery: Add class depending on URL in the best way -

caching - How to check if a url path exists in the service worker cache -

Redirect to a HTTPS version using .htaccess -