神经网络工具箱的使用

天零ヮ冰釋

1. clear
   * y; h* }2 B- g& J
2. load('trainset.mat');                   %加载训练集数据
   & n0 H3 ~: l0 M\" v
3. load('testset.mat');                    %加载测试集数据
   1 k8 q& u4 [! \5 m9 j6 P
4. p1=trainsetdata([1:179],:);+ r( ~6 Q6 {6 v1 ]# q
   
5. p2=trainsetdata([180:300],:);
   2 r: ?$ {; X2 ~9 [
6. p=[p1;p2]';
   0 I3 d! r6 V/ m: K6 e
7. pr=minmax(p);
   : K9 U6 U9 z5 W/ \
8. goal=[ones(1,179),zeros(1,121);zeros(1,179),ones(1,121)];
   7 b7 _$ H5 E) b; I
9. net=newff(pr,[21,160],{'tansig','logsig'}); %隐含层节点和输出层节点9 N% E\" [: c! K) \1 [* R' a0 \
   
10.                                             %取值不太确定，请指教7 |1 ]4 N. Q, @' F6 q! a, e
    
11. net.trainParam.show = 10;% b\" h  {' }5 v5 q
    
12. net.trainParam.lr = 0.05;
    0 X2 e, C$ _+ Y* A
13. net.trainParam.goal = 1e-10;
    5 C7 B\" V' g4 _9 u: ]
14. net.trainParam.epochs = 50000;
    - H: c; K4 b2 \4 Q3 E! [+ g\" ~
15. net = train(net,p,goal);
    3 N& M( r% l1 L2 U; \4 N
16. x=testsetdata';( l  ~& b6 C, d# N) N7 c4 l% @
    
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. \\" P+ U% c$ w3 N0 I4 \% \9 R
4. > In obs_use at 18
   
5. 8 y; R  t3 V. n2 h
6.   In newff>create_network at 127
   
7. ' g7 B2 _- F\\" U5 {. B- A4 ~
8.   In newff at 1026 D$ r0 d8 p' ]7 f: w
   
10.   In test at 9' [- s! y/ B: Y9 y& j$ x
    
12.           See help for NEWFF to update calls to the new argument list.
    
13. \\" E* [+ H  |7 y3 H
14. 
    
15. % `8 ~) F$ G# z8 ?\\" s* E
16. ??? Error using ==> trainlm at 1097 N+ K. l\\" W8 S2 a) O
    
18. Output data size does not match net.outputs{2}.size.
    
19. ) n+ S\\" k$ g* p4 Q5 g  A! j% W# }
20. 
    
21. 6 e: P4 U6 C# I, Y# A2 D
22. Error in ==> network.train at 1074 s$ X# \# z/ u, }8 a& J8 N# f
    
24. [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);
    
25. / @( n) ]. A6 Y) g
26. 0 n% [4 Z5 ?- o( I6 t2 q
    
28. Error in ==> test at 15
    
29. ( f1 E% x  c0 [% ~6 u7 A\\" @
30. net = train(net,p,goal);

Warning: NEWFF used in an obsolete way. <br /> <span style=\\"display:none\\">\\" P+ U% c$ w3 N0 I4 \% \9 R</span> > In obs_use at 18<br /> <span style=\\"display:none\\">8 y; R  t3 V. n2 h</span>   In newff>create_network at 127<br /> <span style=\\"display:none\\">' g7 B2 _- F\\" U5 {. B- A4 ~</span>   In newff at 102<font class=\\"jammer\\">6 D$ r0 d8 p' ]7 f: w</font><br />   In test at 9<font class=\\"jammer\\">' [- s! y/ B: Y9 y& j$ x</font><br />           See help for NEWFF to update calls to the new argument list.<br /> <span style=\\"display:none\\">\\" E* [+ H  |7 y3 H</span> <br /> <span style=\\"display:none\\">% `8 ~) F$ G# z8 ?\\" s* E</span> ??? Error using ==> trainlm at 109<font class=\\"jammer\\">7 N+ K. l\\" W8 S2 a) O</font><br /> Output data size does not match net.outputs{2}.size.<br /> <span style=\\"display:none\\">) n+ S\\" k$ g* p4 Q5 g  A! j% W# }</span> <br /> <span style=\\"display:none\\">6 e: P4 U6 C# I, Y# A2 D</span> Error in ==> network.train at 107<font class=\\"jammer\\">4 s$ X# \# z/ u, }8 a& J8 N# f</font><br /> [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);<br /> <span style=\\"display:none\\">/ @( n) ]. A6 Y) g</span> <font class=\\"jammer\\">0 n% [4 Z5 ?- o( I6 t2 q</font><br /> Error in ==> test at 15<br /> <span style=\\"display:none\\">( f1 E% x  c0 [% ~6 u7 A\\" @</span> net = train(net,p,goal);

jiajinshan110

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