Search code examples
c#equationbackpropagationneural-network

Should I Scale My Equation Output Data?

With this link i implemented a neural network to calculate y=x*x(x is input and y is output) equation.I took 1 unit in input layer,4 unit in hidden layer and 1 unit in output layer. But When I enter for example 2 as input(so desired output is 4) i got output value 0.99999999.... also this happens for all other input numbers bigger than 1(Its output is correct with value between 0 and 1). also I used this link and changed it to solve my equation,but the output was the same!

Solution

  • As Pete Said.I should normalize my output(and input) to get the correct answer.