|
|
Perceptron algorithm in rules
weight change = some small constant ? (target activation - spontaneous output activation) ? input activation
if speak of error instead of the `target activation minus the spontaneous output activation', we have:
weight change = some small constant ? error ? input activation
|