Artificial neural Network (Artificial neural netwroks) Note-delta Rule increment learning

Delta Rule Incremental Learning

Wij (t+1) =wij (t) kit α (Yj-aj (t)) Oi (t)

Type Wij (t+1), Wij (t) represents the ANI of neurons to the Anj at the moment t+1 and the intensity of the time t, Oi (t) for ANI neurons in the output of the time t, YJ for the ideal output of neuron Anj, Aj (t) for the activation state of neuron Anj, α for the given learning rate

This is the most important and widely used delta rule for supervised learning (supervised Learning)

Incremental Learning (Delta Rule Learning) and backward propagation learning (back propagation Learning) can use delta rules to adjust join weights

The theoretical derivation of Delta rules in BP Network is a difficult problem, but it can be achieved by following the formula above