How can create a neural network by hand using trained neural network data in Matlab

for example,following codes create a trained network in Matlab: x = 1:1:50 y = x * 3 net = newff(x,y,5) net = train(net,x,y)

so ,i can get following information from the net: net.IW,net.LW,net.b,and the default transform functions tansig and purelin. Using these information to create a neural network by hand,i find that i can not get the correct result which the function sim() provide. Is there any other important information i ignore when i create a network by hand?

Answer 1

[Greg Heath]

x = 1:1:50; % Semicolon
y = x * 3 ;  % Semicolon
net = newff(x,y,5)    % NO semicolon

inputprocessfunctions  = net.inputs{1}.processFcns    % NO semicolon
outputprocessfunctions = net.outputs{1}.processFcns    % NO semicolon

 help mapminmax

Hope this helps.

Thank you for formally accepting my answer

Greg