神经网络工具箱的使用

天零ヮ冰釋

1. clear
   * `2 n/ \  A# q2 s; O( y& s
2. load('trainset.mat');                   %加载训练集数据
   * p0 c' I- M; e6 v# |$ Q
3. load('testset.mat');                    %加载测试集数据, R1 t$ k\" m2 ]8 ^% k# r3 i
   
4. p1=trainsetdata([1:179],:);
   2 s$ O* w\" b1 o6 f- e  X$ W5 p! t
5. p2=trainsetdata([180:300],:);; I' R1 A$ f* P9 C$ z+ J
   
6. p=[p1;p2]';; ~# h* W6 M0 ]' A
   
7. pr=minmax(p);+ L  e7 X9 p2 s
   
8. goal=[ones(1,179),zeros(1,121);zeros(1,179),ones(1,121)];
   9 ?. m8 S% U/ y8 S1 G+ d
9. net=newff(pr,[21,160],{'tansig','logsig'}); %隐含层节点和输出层节点0 a- t3 Q* J1 A. O- v2 B9 Q& ]
   
10.                                             %取值不太确定，请指教4 f) Z# V7 R( w' ?
    
11. net.trainParam.show = 10;. @! Q$ a0 [5 [) u& W2 m
    
12. net.trainParam.lr = 0.05;
    2 r- O  d& j1 M& ~% X
13. net.trainParam.goal = 1e-10;
    . Z8 g+ ]\" A% C3 }0 N$ A- Q4 i
14. net.trainParam.epochs = 50000;4 d/ U$ A2 G: w5 E
    
15. net = train(net,p,goal);
    9 [1 ]- b  {' I) R
16. x=testsetdata';
    8 @4 g, ~4 t6 F5 N
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. - Y/ u/ Z! W6 D( g# t& K
   
4. > In obs_use at 180 y6 Z3 h. ~- b1 ?: Q7 b
   
6.   In newff>create_network at 127
   
7. ; [) L. }% f, D! i& d
8.   In newff at 102
   
9. & w5 X3 p- Z+ T: ^
10.   In test at 92 L7 {# f( N* }1 A2 G
    
12.           See help for NEWFF to update calls to the new argument list.
    
13. 7 C7 R( v4 @9 h7 T& q$ G
14. # i0 t' i, h# T; J
    
16. ??? Error using ==> trainlm at 109
    
17. 9 \& n# J* `7 S9 l& U7 h2 R, i* \( `1 ?
18. Output data size does not match net.outputs{2}.size.
    
19. + ]- b* ~4 G6 M- R: }
20. 
    
21. ; O7 z9 F9 J1 f4 ~0 P( z9 [
22. Error in ==> network.train at 107
    
23. ; P' X) j3 U4 Z/ o\\" ?
24. [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);
    
25. , A* p1 |) _$ M$ Z& a  C8 @+ i6 d\\" x0 e
26. 4 Z9 A2 b3 B  U. {  c
    
28. Error in ==> test at 157 V5 t* C/ V. ~4 l( k
    
30. net = train(net,p,goal);

Warning: NEWFF used in an obsolete way. <font class=\\"jammer\\">- Y/ u/ Z! W6 D( g# t& K</font><br /> > In obs_use at 18<font class=\\"jammer\\">0 y6 Z3 h. ~- b1 ?: Q7 b</font><br />   In newff>create_network at 127<br /> <span style=\\"display:none\\">; [) L. }% f, D! i& d</span>   In newff at 102<br /> <span style=\\"display:none\\">& w5 X3 p- Z+ T: ^</span>   In test at 9<font class=\\"jammer\\">2 L7 {# f( N* }1 A2 G</font><br />           See help for NEWFF to update calls to the new argument list.<br /> <span style=\\"display:none\\">7 C7 R( v4 @9 h7 T& q$ G</span> <font class=\\"jammer\\"># i0 t' i, h# T; J</font><br /> ??? Error using ==> trainlm at 109<br /> <span style=\\"display:none\\">9 \& n# J* `7 S9 l& U7 h2 R, i* \( `1 ?</span> Output data size does not match net.outputs{2}.size.<br /> <span style=\\"display:none\\">+ ]- b* ~4 G6 M- R: }</span> <br /> <span style=\\"display:none\\">; O7 z9 F9 J1 f4 ~0 P( z9 [</span> Error in ==> network.train at 107<br /> <span style=\\"display:none\\">; P' X) j3 U4 Z/ o\\" ?</span> [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);<br /> <span style=\\"display:none\\">, A* p1 |) _$ M$ Z& a  C8 @+ i6 d\\" x0 e</span> <font class=\\"jammer\\">4 Z9 A2 b3 B  U. {  c</font><br /> Error in ==> test at 15<font class=\\"jammer\\">7 V5 t* C/ V. ~4 l( k</font><br /> net = train(net,p,goal);

jiajinshan110

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