The learning curve denotes the progress of the mistakes, that could be determined in various ways. The motivation to establish a learning curve is that such a curve may indicate whether or not the network is progressing or not. For this, the error really should be normalized i.e. represent a distance measure between the appropriate and the current output of the network. For example, we can take the same design specific, squared error with a prefactor, which we are now also planning to use to derive the backpropagation of error (let be output neurons and other set of output neurons:-