|
Actually updateNeuron is, together with the two other neuron functions, the only really interesting methods in this
program. They are the only methods doing neural computation. The neurons internal signal is set using the linearCombine
function. In mathematics, the line where u is set equals:

As the linear activation function does nothing at all, the next line could be rewritten to y := u, but we included the function
declaration so you could easily change it to another activation function.
The last line in the method does not only update the neuronOutput label, it is also the last described line in this section of the
walkthrough. See you at learning.
|