<![CDATA[AbstractArtificial neural network is a computational algorithm that mimics the workings of nerve cells.All the incoming signal is multiplied by a weight that is on each input, by neuronal cells, allsignals have been multiplied by weights are summed and then added again with bias. Thesum is input to a function (activation function) produces the output of neurons (here used alinear activation function). During the learning process, the weights and biases are alwaysupdated using learning algorithms. if there is an error in the output. For the identificationprocess, the weights are directly weighing input is what is called as a search parameter, theprice of w1, w2, w3 and w4. In the on-line identification, neurons or networks of neurons willalways be 'learned' every input and output data.Keywords: neuro, activation function, learning constants.]]>
Copyrights © 2014
