ADALINE neural network?

[YourSurrogateGod]

I have a rough idea of how they work, but I don't understand some concepts that my book doesn't go into great detail.

There's a formula that simulates learning.
Error = Desired_Output - ADALINE_Output
New_Weight = Weight + Learning_Rate * Error * Input_Value

I'm making an ADALINE network that filters e-mails and I'm not sure what role the Input_Value should play when doing the filtering. Anyone have experience in this field so as to help me elaborate on this? Greatly appreciated.

[yiannakop]

I am not familiar with the specific NN, but in general inputs of NNs are the features of your problem. Have you defined any features for the filtering problem? If yes, the next step is to populate a training set of inputs (feature values) and outputs (desired class labels - e.g. binary decision for rejected\accepted filtering result). If you give some more details I may be able to help you more.
Regards,
Theodore