神经网络工具箱的使用

天零ヮ冰釋

1. clear3 `& }0 k3 C# t' ?1 }0 N# e
   
2. load('trainset.mat');                   %加载训练集数据
   & k6 V0 p& C\" w
3. load('testset.mat');                    %加载测试集数据7 }% P: I) _  {- q( V
   
4. p1=trainsetdata([1:179],:);
   - ?# d3 Q3 H# m' @! b
5. p2=trainsetdata([180:300],:);
   ' M  j( M\" p8 \2 Y$ r
6. p=[p1;p2]';
   ( V: s. z% ~- z1 m( v
7. pr=minmax(p);
   & Z8 {+ I8 t* ?+ B
8. goal=[ones(1,179),zeros(1,121);zeros(1,179),ones(1,121)];; g9 a. z: D. ~8 c+ x/ x
   
9. net=newff(pr,[21,160],{'tansig','logsig'}); %隐含层节点和输出层节点
   % e$ G$ e4 c8 A( N1 y
10.                                             %取值不太确定，请指教
    4 M1 m6 V5 w, X( z7 j
11. net.trainParam.show = 10;
    * B6 x* u+ c0 P; f3 c3 K2 b$ T- r$ V\" R
12. net.trainParam.lr = 0.05;
    7 y/ a2 m1 k5 X* ^
13. net.trainParam.goal = 1e-10;
    # v! Y% h! J6 b\" s' M' F
14. net.trainParam.epochs = 50000;
    : F* S7 \3 y/ I. `1 D
15. net = train(net,p,goal);* i# R9 a& S; `1 ^8 o
    
16. x=testsetdata';
    / N1 l2 V% X- E6 s9 k# Z
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 A# J0 X% y5 L4 C/ r) c# @/ A0 H
   
4. > In obs_use at 18
   
5. ( S% I\\" ]. ?) u0 M% W1 ]
6.   In newff>create_network at 127
   
7. $ @1 K! ?$ C0 Q
8.   In newff at 102
   
9. 8 p* R* X9 q  m# P
10.   In test at 9
    
11. 5 ^  u) ~8 i/ H
12.           See help for NEWFF to update calls to the new argument list.\\" U: w0 ?1 l8 {2 u9 E
    
14. 
    
15.   Q# F- g, ?7 \: t; U
16. ??? Error using ==> trainlm at 109
    
17. - I5 r( _) F; u( E4 ]  _
18. Output data size does not match net.outputs{2}.size.
    
19. ; x; o\\" k/ B: y5 g: \6 t
20. 1 ]( }+ N/ M. z\\" Z; }# W3 O  L% C! f
    
22. Error in ==> network.train at 107& w* B7 C5 i5 z6 j7 C8 ^
    
24. [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);! {1 @1 o2 B( e3 m* J
    
26. 
    
27. & \5 k' p2 X8 G* U, Z
28. Error in ==> test at 15
    
29. / j9 L1 C0 g+ c# j0 q! ]1 w2 F7 a9 {
30. net = train(net,p,goal);

Warning: NEWFF used in an obsolete way. <font class=\\"jammer\\">3 A# J0 X% y5 L4 C/ r) c# @/ A0 H</font><br /> > In obs_use at 18<br /> <span style=\\"display:none\\">( S% I\\" ]. ?) u0 M% W1 ]</span>   In newff>create_network at 127<br /> <span style=\\"display:none\\">$ @1 K! ?$ C0 Q</span>   In newff at 102<br /> <span style=\\"display:none\\">8 p* R* X9 q  m# P</span>   In test at 9<br /> <span style=\\"display:none\\">5 ^  u) ~8 i/ H</span>           See help for NEWFF to update calls to the new argument list.<font class=\\"jammer\\">\\" U: w0 ?1 l8 {2 u9 E</font><br /> <br /> <span style=\\"display:none\\">  Q# F- g, ?7 \: t; U</span> ??? Error using ==> trainlm at 109<br /> <span style=\\"display:none\\">- I5 r( _) F; u( E4 ]  _</span> Output data size does not match net.outputs{2}.size.<br /> <span style=\\"display:none\\">; x; o\\" k/ B: y5 g: \6 t</span> <font class=\\"jammer\\">1 ]( }+ N/ M. z\\" Z; }# W3 O  L% C! f</font><br /> Error in ==> network.train at 107<font class=\\"jammer\\">& w* B7 C5 i5 z6 j7 C8 ^</font><br /> [net,tr] = feval(net.trainFcn,net,X,T,Xi,Ai,EW,net.trainParam);<font class=\\"jammer\\">! {1 @1 o2 B( e3 m* J</font><br /> <br /> <span style=\\"display:none\\">& \5 k' p2 X8 G* U, Z</span> Error in ==> test at 15<br /> <span style=\\"display:none\\">/ j9 L1 C0 g+ c# j0 q! ]1 w2 F7 a9 {</span> net = train(net,p,goal);

jiajinshan110

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