where output represents the neural network output in the (0,1) range, data range represents the range of the original attribute values on the non normalized scale, and minimum represents the smallest attribute values on the non normalized scale. for example, suppose that the stock prices ranged from $20 to $30 and that the network output was 0.69 . then the predicted stock price in three months is
prediction = output (data range) + minimum = 0.69 ($10) + $20 = $26.90