反馈型神经网络理论及MATLAB实现的几个程序

frogfish

摘自：神经网络理论与MATLAB7实现

Elman网络设计

1. p1 = sin(1:20);
   
2. p2 = sin(1:20)*2;
   
3. t1 = ones(1,20);
   
4. t2 = ones(1,20)*2;
   
5. %产生训练样本p和t
   
6. p = [p1 p2 p1 p2];
   
7. t = [t1 t2 t1 t2];
   
8. Pseq = con2seq(p);
   
9. Tseq = con2seq(t);
   
10. R = 1; % 输入元素的数目为1
    
11. S2 = 1; % 输出层的神经元个数为1
    
12. S1 = 10; %中间层有10个神经元
    
13. net=newelm([-2,2],[S1,S2],{'tansig','purelin'});
    
14. %设定网络训练次数
    
15. net.trainParam.epochs=300;
    
16. net=train(net,Pseq,Tseq);
    
17. y=sim(net,Pseq);
    
18. figure;
    
19. plot(t5,cat(2,y{:}),t5,cat(2,Tseq{:}),'b--');
    
20. %利用新的信号来测试网络
    
21. p3 = sin(1:20)*1.6;
    
22. t3 = ones(1,20)*1.6;
    
23. p4 = sin(1:20)*1.2;
    
24. t4 = ones(1,20)*1.2;
    
25. %产生测试样本pg和tg
    
26. pg = [p3 p4 p3 p4];
    
27. tg = [t3 t4 t3 t4];
    
28. pgseq = con2seq(pg);
    
29. a = sim(net,pgseq);
    
30. figure;
    
31. plot(t5,cat(2,a{:}),t5,tg,'b--');

复制代码

[ 本帖最后由 frogfish 于 2007-6-25 03:20 编辑 ]

frogfish

1. P=[0.4413 0.4707 0.6953 0.8133 0.4379 0.4677 0.6981 0.8002 0.4517 0.4725 0.7006 0.8201;
   
2. 0.4379 0.4677 0.6981 0.8002 0.4517 0.4725 0.7006 0.8201 0.4557 0.4790 0.7019 0.8211;
   
3. 0.4517 0.4725 0.7006 0.8201 0.4557 0.4790 0.7019 0.8211 0.4601 0.4811 0.7101 0.8298;]'
   
4. T=[0.4557 0.4790 0.7019 0.8211;
   
5. 0.4601 0.4811 0.7101 0.8298;
   
6. 0.4612 0.4845 0.7188 0.8312]';
   
7. threshold=[0 1;0 1;0 1;0 1;0 1;0 1;0 1;0 1;0 1;0 1;0 1;0 1];
   
8. a=[11 17 23];
   
9. for i=1:3
   
10.     net=newelm(threshold,[a(i),4],{'tansig','purelin'});
    
11.     net.trainParam.epochs=1000;
    
12.     net=init(net);
    
13.     net=train(net,P,T);
    
14.     y=sim(net,p_test);
    
15.     error(i,:)=y'-t;
    
16. end
    
17. hold off;
    
18. plot(1:4,error(1,:));
    
19. hold on;
    
20. plot(1:4,error(2,:),'-.');
    
21. hold on;
    
22. plot(1:4,error(3,:),'--');
    
23. hold off;

复制代码

frogfish

1. one=[-1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1  -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1];
   
2. two=[1 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 1 1 -1 -1];
   
3. T=[one;two]';
   
4. net=newhop(T);
   
5. no2={[1 1 1 -1 1 1 -1 1 -1 -1  1 1 1 1 1 1 1 1 -1 -1  -1 1 -1 1 -1 -1 1 1 -1 -1 -1 -1 1 -1 1 -1 1 1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 1 1 -1 -1  1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 -1 1 1 1 1 -1 -1 1 1 -1 1 1 1 1 1 -1 -1]'};
   
6. tu2=sim(net,{1,5},{},no2);
   
7. tu2{3}'

复制代码

frogfish

1. function c=bsb(x,beta,multi)
   
2. 
   
3. % function c=bsb(x,beta)
   
4. %
   
5. %  This m-file duplicates the Brain State in a Box Experiment.
   
6. %    x     - input vector
   
7. %    beta  - feedback factor
   
8. %    c     - number of iterations required for convergence
   
9. %   
   
10. % Hugh Pasika 1997
    
11. 
    
12. hold on
    
13. flag=0;   x=x(:);   c=2; %c is a general purpose counter
    
14. W=[.035 -.005; -.005 .035];
    
15. 
    
16. set(gca,'YLim',[-1 1]);      set(gca,'XLim',[-1 1])    % set axes
    
17. 
    
18. plot(x(1),x(2),'ob')         % plot first point   
    
19. orig=x';
    
20. plot([0,0],[1,-1],'-');      plot([1,-1],[0,0],'-')     % plot center lines
    
21. set(gca,'YTick',[-1  1 ]);   set(gca,'XTick',[-1  1 ])  % label plot
    
22. 
    
23. while flag < 1,
    
24. y=x+beta*W*x;
    
25. x=(y(:,:) < -1 )*(-1) + (y(:,:)>1) + (y(:,:) > -1 & y(:,:) < 1).*y;
    
26. u(c,:)=x';
    
27. c=c+1;
    
28. if u(c-1,:) == u(c-2,:),
    
29.    flag=10;
    
30.    c=c-3;
    
31. end
    
32. end
    
33. u=u(2:c+1,:);
    
34. orig
    
35. plot([orig(1,1) u(1,1)], [orig(1,2) u(1,2)],'-b')
    
36. plot(u(:,1), u(:,2),'ob')
    
37. plot(u(:,1), u(:,2),'-b')
    
38. drawnow
    
39. fprintf(1,'It took %g iteration for a stable point to be reached.\n\n',c);
    
40. set(gca,'Box','on')
    
41. hold off

复制代码

adiko

谢谢~~:loveliness:

冬冬星LWWL

楼主辛苦:handshake

zhuxiaoxun

请问第一个程序的t5变量在哪定义的？

leei8210

谢谢了！楼主辛苦了！

deepedlver

原帖由 zhuxiaoxun 于 2009-3-18 22:29 发表

                               
登录/注册后可看大图

请问第一个程序的t5变量在哪定义的？

同问，新学的不太明白

锦在熙

为什么该程序一运行就显示“??? Error using ==> feedback at 72
Not enough input arguments. ”