Shopping Podder - the Best of Computer Postings! :: View topic

Berdakh Abibullaev · Guest

Hello ,

I am learning Neural Networks and very new in this area. Thus, I want
to train one neuron perceptron with simple perceptron rule. There are
four inputs p(1, Smile

,p(2,

,p(3,

,p(4,

and two target outputs as
t[1,1,0,0] . I am using hardlim transfer function. Can you please
check my matlab code below and correct my mistakes.

Thanks!
---------------------------------------
clear all; close all;
% set input vectors
p = [-1,1; -1,-1; 0,0; 1,0];
% define target values
t = [1,1,0,0];
% set initial weights and bias
w = [0,0]; b = 0;
ii = 0; jj = 0; nn = 1;
aa = rand(1,4);

% start training the network
while jj == 0

fprintf('\nProcessing...\n')

if nn > 4
nn = 1;
end
pp = nn;

n = w*p(nn,1:2)' + b;

% here we threshold using transfer function in our case a =
harlim(n)
if n >= 0
a = 1;
elseif n < 0
a = 0;
end

% now calculate the error e = (t-a)
if pp == 1
e = t(1,1)-a;
aa(1,pp) = e;
elseif pp ==2
e = t(1,2)-a;
aa(1,pp) = e;
elseif pp == 3
e = t(1,3)-a;
aa(1,pp) = e;
elseif pp == 4
e = t(1,4)-a;
aa(1,pp) = e;
end

% following the step we update weight and bias
w = w+e*p(nn,1:2)+b;
b = b + e;

if aa(1, Smile

== t(1,

disp('Completed')
jj = 1;
end

nn = nn + 1;

end


www.ShoppingPodder.com Leading Computer Shopping, News and information	Part of the Identityscape.com network... getxfactor.com jmoodmusic.com smartbusinesschoices.com mintdepot.com lowfaresalways.com evangelicalview.com shoppingpodder.com soproudlywehail.com webnews.ws currenthumor.com

Shopping Podder - the Best of Computer Postings! Forum Index -> Computer Artificial Intelligence - Neural Nets
Page 1 of 1