神经网络工具箱的使用

天零ヮ冰釋

1. clear
   2 q& m+ }6 s# ^) c
2. load('trainset.mat');                   %加载训练集数据
   ! h, d( S+ j/ k' M
3. load('testset.mat');                    %加载测试集数据
   + f- Y0 g9 ~( N+ U7 D$ y
4. p1=trainsetdata([1:179],:);
   6 J4 `9 e\" l3 V: Y( R- m1 n$ n
5. p2=trainsetdata([180:300],:);0 v7 }) \2 w4 E( k
   
6. p=[p1;p2]';
   - _) e- t+ R8 o: P: C! }, R* h
7. pr=minmax(p);& J- B. c) U8 W
   
8. goal=[ones(1,179),zeros(1,121);zeros(1,179),ones(1,121)];0 _0 J9 P! Z. I. G
   
9. net=newff(pr,[21,160],{'tansig','logsig'}); %隐含层节点和输出层节点' J0 w( {) c7 m/ O1 @$ N  X
   
10.                                             %取值不太确定，请指教
    / ~9 f. m, m9 w0 ]. N6 P# |3 o9 D
11. net.trainParam.show = 10;
    ; i) x/ _- p# A
12. net.trainParam.lr = 0.05;
    , o: G0 ^5 K! b8 w5 c\" h3 [
13. net.trainParam.goal = 1e-10;- p\" u7 P; b( D4 i6 {' L
    
14. net.trainParam.epochs = 50000;
    5 \* B3 j- J; U% Z
15. net = train(net,p,goal);4 ]  r& G0 _7 O2 e: H
    
16. x=testsetdata';. x) }/ p\" \. K& C  D8 r
    
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. + J/ P& }2 B  M& ^* Q4 W% l
   
4. > In obs_use at 18
   
5.   k) @1 A* }( a. a! m
6.   In newff>create_network at 127% k/ G' f7 W/ i2 u\\" D
   
8.   In newff at 102
   
9. . p1 R  U% O1 O% W8 q6 ]( [
10.   In test at 9
    
11. & `9 {5 K+ ^6 ?# X1 G
12.           See help for NEWFF to update calls to the new argument list.
    
13. ; E: s\\" t( ]* {/ `9 |
14. 
    
15. , s9 g9 A: J6 ]8 R# g/ d6 u+ F
16. ??? Error using ==> trainlm at 109
    
17. ( y# w5 c4 X/ F6 x# H
18. Output data size does not match net.outputs{2}.size.
    
19. 6 W/ W2 }% E$ F
20. 9 v* |2 o& g) F- h5 g1 X% r2 w
    
22. Error in ==> network.train at 107
    
23. 7 G, s) G' m/ }7 U3 l# w0 E! I
24. [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);. a) f2 h0 i( J1 c/ x% n& j* ^
    
26. 8 ~/ l* t1 r  }$ O2 Q5 ?: a
    
28. Error in ==> test at 152 ^. L4 v2 n2 X; }5 x# d$ y
    
30. net = train(net,p,goal);

Warning: NEWFF used in an obsolete way. <font class=\\"jammer\\">+ J/ P& }2 B  M& ^* Q4 W% l</font><br /> > In obs_use at 18<br /> <span style=\\"display:none\\">  k) @1 A* }( a. a! m</span>   In newff>create_network at 127<font class=\\"jammer\\">% k/ G' f7 W/ i2 u\\" D</font><br />   In newff at 102<br /> <span style=\\"display:none\\">. p1 R  U% O1 O% W8 q6 ]( [</span>   In test at 9<br /> <span style=\\"display:none\\">& `9 {5 K+ ^6 ?# X1 G</span>           See help for NEWFF to update calls to the new argument list.<br /> <span style=\\"display:none\\">; E: s\\" t( ]* {/ `9 |</span> <br /> <span style=\\"display:none\\">, s9 g9 A: J6 ]8 R# g/ d6 u+ F</span> ??? Error using ==> trainlm at 109<br /> <span style=\\"display:none\\">( y# w5 c4 X/ F6 x# H</span> Output data size does not match net.outputs{2}.size.<br /> <span style=\\"display:none\\">6 W/ W2 }% E$ F</span> <font class=\\"jammer\\">9 v* |2 o& g) F- h5 g1 X% r2 w</font><br /> Error in ==> network.train at 107<br /> <span style=\\"display:none\\">7 G, s) G' m/ }7 U3 l# w0 E! I</span> [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);<font class=\\"jammer\\">. a) f2 h0 i( J1 c/ x% n& j* ^</font><br /> <font class=\\"jammer\\">8 ~/ l* t1 r  }$ O2 Q5 ?: a</font><br /> Error in ==> test at 15<font class=\\"jammer\\">2 ^. L4 v2 n2 X; }5 x# d$ y</font><br /> net = train(net,p,goal);

jiajinshan110

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