Step 3:
clear;
clc;
close all;
mynetwork=feedforwardnet([7,7]); % Create the network
mynetwork.layers{1:2}.transferFcn=
'logsig'; % Change the activation functions for the hidden layers (from default tansig to logistic logsig)
mynetwork.layers{3}.transferFcn=
'tansig'; % Change the activation function for the output layer
(from pure linear to tansig)
Change the learning rule
to stochastic gradient descent (or to another one of the ones shown in
the list during lecture). Some of these rules are faster than
others, and some are more accurate. Some are designed for smaller
networks and some for larger complex networks. The stochastic
gradient descent training rule is a good compromise among the many
training rules and can be used in most cases.
mynetwork.trainFcn = 'traingd';