Backprop Network
Backprop
is a standard form of supervised learning. It is perhaps the best known
and most popular means of training neural networks.
An example which steps through the process of creating and trainining a backprop network is in the examples page.
Creation Dialog
Number of Layers
Number of Neurons
Neuron Type
Tanh
Logistic
Linear
After selecting Insert > Insert Network > Backrpop, a dialog will appear asking that you specify a topology, which means how many layers, and how many neurons per layer (from bottom to top layer).
Right Click Menu
Edit/Train Backprop
Rename
Remove Network
View/Edit Data
Training
Training a network involves specifying input data, target data, and then running the algorithm. This process is covered here.
Parameters
Learning Rate: A standard learning rate. This determines how quickly synapses change.
Momentum: This scales the rate of weight change by the amount a given weight changed on the previous time step. This speeds up learning and prevents oscillations. Momentum should be between 0 and 1; 0.9 is a common value.