In the neural network classification algorithm, Tiberius predicts positive numbers for one class and negative numbers for the other class. Zero is known as the 'decision boundary'.
The best solution is the one with the minimum 'cost', which can be considered a monetary value. It may not matter if we get 9 non-responders to a mail campaign that costs $1 per mailing if the 10th will repspond and result in a $11 sale. We have still made money.
This table allows costs to be entered so that Tiberius can find the appropriate decision boundary in order to minimise the cost.
