ADALINE

Adaline the model is a form of neural network. The name stood for ADAptive LInear neuron lost as neural networks later in popularity, the acronym was also translated as ADAptive LINear element.

The Adaline model was proposed in 1960 by Bernard Widrow and Marcian Edward Hoff and forms a perceptron in which the learning algorithm is also the activation of an output unit as information is available. The picture shows the structure of an ADALINE. The learning algorithm adapts the weights of the input vector and the bias ( bias) input.

A learning control Widrow and Hoff proposed the so-called delta rule (also known as a Widrow-Hoff rule, or Least Mean Squares ) algorithm. This causes a different behavior of the artificial neural network by building more stable knowledge, since once learned knowledge is less easily forgotten by the modified learning rule.

• Neuro computer science