How can I get MSE and normalized MSE both as performance function when fitting Feed forward neural network?

Im fitting a feedforward neural network with 8 input parameters and 1 output parameter.i want to use MSE and normalized MSE both to measure the performance.

Answer 1

[Greg Heath]

Divide MSE by the mean variance of the target rows (MSE of the constant output model)

 MSEref = mean(var(target',1))

 NMSE = mse(target-output)/MSEref

 Rsquare = 1 - NMSE

Hope this helps.

Thank you for formally accepting my answer

Greg

[Avanthi Saumyamala]

Thank you very much Greg

I saw from one of your answer MSEref computed as follows.

The average biased (e.g., divide by N) target variance is

 MSE00 = mean(var(t'),1)

Is this the same or what is the correct answer?

Thank you

Avanthi

[Greg Heath]

Same