神经网络工具箱的使用

天零ヮ冰釋

1. clear
   8 g+ i, R$ E( |7 F) v
2. load('trainset.mat');                   %加载训练集数据
     K7 d/ i( W. y+ n! S1 o
3. load('testset.mat');                    %加载测试集数据
   6 g% H9 u5 w' h. ]4 K# H, Z( U
4. p1=trainsetdata([1:179],:);
   * N# D' d0 B1 C2 G: P
5. p2=trainsetdata([180:300],:);- `3 C5 D$ O' ~1 j: A
   
6. p=[p1;p2]';  \1 a; w% S/ c/ j+ w( U5 }
   
7. pr=minmax(p);
   % A( U. H1 J( c7 V
8. goal=[ones(1,179),zeros(1,121);zeros(1,179),ones(1,121)];
   ! z& p. }  r7 ~3 k8 Q
9. net=newff(pr,[21,160],{'tansig','logsig'}); %隐含层节点和输出层节点1 [3 ?% w# m. c2 g9 e2 s
   
10.                                             %取值不太确定，请指教
    , _3 H  t9 f! y5 b1 W
11. net.trainParam.show = 10;5 Y( u, M4 _) Q8 [4 J7 ^
    
12. net.trainParam.lr = 0.05;0 F; i\" U/ w/ d3 `+ F/ b, F6 k. N9 m
    
13. net.trainParam.goal = 1e-10;6 t) @- j9 P; I; i# `/ T4 }/ @
    
14. net.trainParam.epochs = 50000;
    . c% i$ A& R8 @+ V
15. net = train(net,p,goal);
    5 X$ W! V* s7 k. }) d
16. x=testsetdata';
    & l\" Z9 Y! u4 Y: {& y
17. y0=sim(net,p)

复制代码

下面是运行结果，请各位高手帮忙改改啊，主要是节点数那块儿不知道该怎么取值

2. Warning: NEWFF used in an obsolete way.
   
3. ; L7 r9 ~' M- X0 v. f! n
4. > In obs_use at 18
   
5.   k4 m* r8 Y\\" ~4 c: f- }
6.   In newff>create_network at 127
   
7. % P4 b. q  N! P2 j, ^- O% |* R
8.   In newff at 102# u  Z; E: P* k6 C
   
10.   In test at 9
    
11. 7 ^0 K% c. v* k: L6 V& A- a# F
12.           See help for NEWFF to update calls to the new argument list.  ~4 P, ]\\" V. w
    
14. 
    
15. \\" r- R\\" ~  v, E; T* `
16. ??? Error using ==> trainlm at 109
    
17. ' J# c; @9 M+ U; `% n3 k
18. Output data size does not match net.outputs{2}.size.
    
19. 2 g5 d) E6 c7 u( |5 l( q
20. 6 Z* S9 n4 [, l9 ], u/ d+ ~8 k
    
22. Error in ==> network.train at 1070 O( V. `9 V6 q
    
24. [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);5 U- S) o, T- ^) u2 y
    
26. 
    
27. 3 w. I) u1 S$ _
28. Error in ==> test at 15
    
29. % P' x- @, D4 p# @+ `* F
30. net = train(net,p,goal);

Warning: NEWFF used in an obsolete way. <br /> <span style=\\"display:none\\">; L7 r9 ~' M- X0 v. f! n</span> > In obs_use at 18<br /> <span style=\\"display:none\\">  k4 m* r8 Y\\" ~4 c: f- }</span>   In newff>create_network at 127<br /> <span style=\\"display:none\\">% P4 b. q  N! P2 j, ^- O% |* R</span>   In newff at 102<font class=\\"jammer\\"># u  Z; E: P* k6 C</font><br />   In test at 9<br /> <span style=\\"display:none\\">7 ^0 K% c. v* k: L6 V& A- a# F</span>           See help for NEWFF to update calls to the new argument list.<font class=\\"jammer\\">  ~4 P, ]\\" V. w</font><br /> <br /> <span style=\\"display:none\\">\\" r- R\\" ~  v, E; T* `</span> ??? Error using ==> trainlm at 109<br /> <span style=\\"display:none\\">' J# c; @9 M+ U; `% n3 k</span> Output data size does not match net.outputs{2}.size.<br /> <span style=\\"display:none\\">2 g5 d) E6 c7 u( |5 l( q</span> <font class=\\"jammer\\">6 Z* S9 n4 [, l9 ], u/ d+ ~8 k</font><br /> Error in ==> network.train at 107<font class=\\"jammer\\">0 O( V. `9 V6 q</font><br /> [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);<font class=\\"jammer\\">5 U- S) o, T- ^) u2 y</font><br /> <br /> <span style=\\"display:none\\">3 w. I) u1 S$ _</span> Error in ==> test at 15<br /> <span style=\\"display:none\\">% P' x- @, D4 p# @+ `* F</span> net = train(net,p,goal);

jiajinshan110

..................