BP神经网络程序问题求助

T-Eric

这是我从书上打下来的代码，因为书的运行软件版本比较低，我的是Malab2013a的版本，所以我修改了部分程序，但运行出来有问题。请哪位高手帮忙指教和修正一下。谢谢！！
+ P3 V( \" g. {7 W5 h
%原始数据输入
. Y; P( g* P4 I' I9 y( U/ Jclc
sqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
4 H) P" J" I9 [' q; C$ n    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
sqjdcs=[0.6,0.75,0.85,0.90,1.05,1.35,1.45,1.60,1.70,1.85,2.15,2.20,2.25,2.35,...
# Q# s9 C' K. d' G7 t5 `    2.5,2.6,2.7,2.85,2.95,3.10];
0 y1 C3 |: I7 u; ~' x7 I" q( Gsqglmj=[0.09,0.11,0.11,0.14,0.20,0.23,0.23,0.32,0.32,0.34,0.36,0.36,0.38,0.49,...
- M. R; I; Y% l0 K# i: T    0.56,0.59,0.59,0.67,0.69,0.79];
glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
    22598,25107,33442,36836,40548,42927,43462];
glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    13320,16762,18673,20724,20803,21804];
/ }/ Q0 ~- N0 G+ u: m) h% w( Xp=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];
/ s. F$ U3 @' ~: i% T
%数据归一化
# P( Y% W4 {3 d4 Z, O" u) Y# m[pn,ps1]=mapminmax(p);
! p3 v! X1 P+ k7 a& N* q& m5 |[tn,ps2]=mapminmax(t);
4 q5 y; g! S% X6 q3 Edx=[-1,1;-1,1;-1,1];
5 v% R+ o' r7 o( I0 H1 [3 ~) n1 d
%BP网络训练
0 C) K! b  S! u* z, Znet=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
5 _/ L4 }. N& q  H/ G* S( C' ]7 cnet.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
net=train(net,pn,tn);

%利用原数据对BP网络仿真
an=sim(net,pn);
* j# a' N- X) b- |6 da=mapminmax('reverse',an,ps2);
, e6 R1 H0 m) E3 S( r
5 I" k: y) }' \* ]%仿真结果与原数据对比测试
x=1990:2009;
& J  l+ P' V8 q# a3 x" unewk=a(1,:);
' m4 Y# ~- ~* ~' a+ s: }; tnewh=a(2,:);
7 _4 ?1 d! f* K% K3 b: \$ q* J8 [- W! \figure(2);
! W* X- }) V8 s! G& S) T) Esubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
legend('网络输出客运量','实际客运量');
xlabel('年份');ylabel('货运量、万人');
title('运用工具箱客运量学习和测试对比图');
, T+ |  Q" N. g$ f4 [subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
legend('网络输出货运量','实际货运量');
xlabel('年份');ylabel('货运量、万吨');
# J; H% q, L0 H( M; A3 O/ utitle('运用工具箱货运量学习和测试对比图');
- Z) E- {3 ]1 E& U. w( D( M
%新数据仿真
- x- h" x* x# tpnew=[73.39,75.55
    3.9635,4.0975
    0.9880,1.0268];
pnewn=mapminmax('apply',pnew,ps1);
# E  N6 W  n+ F5 canewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)

0 d) R+ U- G) }

T-Eric

狼之魂汪洋 发表于 2014-8-6 14:00
. O) C2 }3 D7 [1 _" o! J0 D你这例题是在什么书上找的？

就是那本《matalb在数学建模中的应用》

狼之魂汪洋

你这例题是在什么书上找的？

且生

因为样本比较少，所以训练函数没有选会调节学习率的

且生

1. %BP网络训练/ I5 z; I, j: ?! i! c: {' x
   
2. net=newff(p0,t0,6,{'tansig'},'traingd'); 1 m4 |0 K5 a3 u! s
   
3. net.trainParam.show=1000;
   2 a: G  F# }4 V9 a- D
4. net.trainParam.Lr=0.05;* ]9 X$ Z2 C1 c4 H& ]+ o, ~) Y
   
5. net.trainParam.epochs=2000;
   & _8 V3 q. n  i8 s3 F; z( K
6. net.trainParam.goal=0.65*10^(-4);# i/ m+ q8 T; ^0 h
   
7. net=train(net,pn,tn);

复制代码

我只改了这里面的，
一、改成单隐含层的，6个节点
二、训练函数改成梯度下降BP算法 traingd
三、迭代次数改成2000
上面的参数是自己试的，我也不知道为什么。
% f) b/ J- z4 O' C5 E, R/ `/ a由于这玩意儿比较不靠谱，楼主多运行几次就能找到拟合的比较好的网络。
8 F' O7 F6 L0 M1 b关键问题是有没有过拟合我也不知道，等大神来解答吧

T-Eric

且生 发表于 2014-7-21 18:22
忽略我上面那个，改动的地方在18，21，25行……求批

确实改善了很多，很是感谢。但效果还是不太理想，而且为何只学习了几十次就停了呢？即使我修改了目标精度。

且生

忽略我上面那个，改动的地方在18，21，25行……求批

1. clc( h\" z1 u$ v' ~7 r9 _9 _\" h
   
2. sqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,.... d6 J$ w5 H& @! G, ~% l0 v
   
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
   5 \. r7 F8 `- h& l, C2 l& v) ^; [2 @
4. sqjdcs=[0.6,0.75,0.85,0.90,1.05,1.35,1.45,1.60,1.70,1.85,2.15,2.20,2.25,2.35,...
   8 v) n, e3 c# Q. s6 [\" C: H\" K
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   ( _5 I; ^3 K$ y$ w6 y
6. sqglmj=[0.09,0.11,0.11,0.14,0.20,0.23,0.23,0.32,0.32,0.34,0.36,0.36,0.38,0.49,...
   4 w) @/ C5 c% t* @; b7 \9 a
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   9 O4 z\" K1 J/ V4 D& A# ]
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
   : D7 _: g: o$ Z: l
9.      22598,25107,33442,36836,40548,42927,43462];
   ; l+ `  T- d  T- I2 u
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    $ E  h\" ]4 v& E, I- A\" [2 i7 w
11.      13320,16762,18673,20724,20803,21804];) h5 w6 Y8 t$ Z2 f2 p7 M
    
12. p=[sqrs;sqjdcs;sqglmj];
    0 v+ h% Y' D/ {( R* R2 z( l( g
13. t=[glkyl;glhyl];, d  i9 ~) u5 K8 u) e) B
    
14. %数据归一化) j  ?: D+ W9 c# l$ _
    
15. [pn,ps1]=mapminmax(p);
    * n$ F' J% ^& i) f
16. [tn,ps2]=mapminmax(t);
    3 U! ]* ^6 L# B1 }% X
17. %dx=[-1,1;-1,1;-1,1];
    0 i9 v' G0 |% g
18. p0=minmax(pn);t0=minmax(tn);
    . i$ n/ i% o8 V- @0 F\" R
19. 
    / ]4 Y/ z& C9 O) w6 T, Z2 v
20. %BP网络训练+ b% j/ l\" R8 r$ x
    
21. net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); : C) z$ f1 D2 N, x; }) F
    
22. net.trainParam.show=1000;\" L$ G: u2 b9 U. Q
    
23. net.trainParam.Lr=0.05;4 x0 V3 k* g) f, W' u) {) G
    
24. net.trainParam.epochs=50000;' X\" c# N$ \) c, J\" q
    
25. net.trainParam.goal=0.65*10^(-5);
    5 G  l  @; ^* ~! A, K  K$ g* d
26. net=train(net,pn,tn);/ b0 l0 [5 k+ i2 U; m; k
    
27. 
    : A; \* J- e/ w9 J( q
28. %利用原数据对BP网络仿真
    * n' G! `; G! C6 K6 i% _
29. an=sim(net,pn);5 I& r! ?. J* ]* X\" b6 |
    
30. a=mapminmax('reverse',an,ps2);
    1 V+ y4 d( V+ V
31. 8 Y+ ~: E- l2 n9 c( O# }
    
32. %仿真结果与原数据对比测试 5 i: e2 u% u* x  R) ?; \8 ?
    
33. x=1990:2009;\" d' F5 A: Z5 Z  }3 J' @- D8 |
    
34. newk=a(1,:);
    % b, W4 }4 B: G2 h
35. newh=a(2,:);
    , _- V8 y5 I' I& R; ^
36. figure(2);
    7 V; @$ k1 V5 L1 A$ h
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
    * z% ~( {+ ^+ Y( P
38. legend('网络输出客运量','实际客运量');# r' a- [6 j3 O8 W\" w5 [6 Z
    
39. xlabel('年份');ylabel('货运量、万人');
    . H% t- P/ W! P3 P. Y
40. title('运用工具箱客运量学习和测试对比图');
    # x. N4 J1 r1 B9 r& R! ?; Z\" p
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
    ! z+ z7 N- q2 m* C: j
42. legend('网络输出货运量','实际货运量');- y' E  v0 X, U. L+ B1 Q
    
43. xlabel('年份');ylabel('货运量、万吨');/ z, @, v% m: a* ^, t3 `
    
44. title('运用工具箱货运量学习和测试对比图');
    : s. R$ a$ j& }- l; J& q0 Q6 o( I
45. 4 {4 A+ ~( q$ Z' _1 @, H# t
    
46. %新数据仿真
    ; F8 ^+ g$ @* p
47. pnew=[73.39,75.552 i% F+ U/ T' l7 r: Y3 {1 I8 O! l
    
48.      3.9635,4.09755 |$ |3 a4 b* ?* ^
    
49.      0.9880,1.0268];( \; r6 y3 V( w# p( O
    
50. pnewn=mapminmax('apply',pnew,ps1);
    * E3 Q6 n4 m8 C8 {! X5 b. T
51. anewn=sim(net,pnewn);
    ' x; e3 t. R  e3 c
52. anew=mapminmax('reverse',anewn,ps2)

复制代码

且生

1. clc
   ) e' x7 n& P. X! J
2. sqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
   ' K6 x% h, v+ g- }& ?$ b  m
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];& V4 B2 B. G. I2 k, F
   
4. sqjdcs=[0.6,0.75,0.85,0.90,1.05,1.35,1.45,1.60,1.70,1.85,2.15,2.20,2.25,2.35,...& C' x$ i2 w( n6 ]- w, O$ l9 u
   
5.     2.5,2.6,2.7,2.85,2.95,3.10];, ^9 m\" k: u4 I( _% h
   
6. sqglmj=[0.09,0.11,0.11,0.14,0.20,0.23,0.23,0.32,0.32,0.34,0.36,0.36,0.38,0.49,...
   7 Z( Z1 q# L$ Z
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   8 |  ]& v5 n+ \. K; j6 c- ^
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...) j- v- R3 u, S! p8 L4 F7 s
   
9.      22598,25107,33442,36836,40548,42927,43462];' s1 v- L! ]' R/ B- f# u2 }) C
   
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    0 P; z( X/ c) h# A' d4 e
11.      13320,16762,18673,20724,20803,21804];
    ! A8 F/ b/ i  i( L9 I
12. p=[sqrs;sqjdcs;sqglmj];
    3 M/ X& a9 Y8 x\" L5 o3 ]  l
13. t=[glkyl;glhyl];; X& U0 s( M& w8 J4 M, Q* z+ a- X) A
    
14. %数据归一化3 Z. O( |3 W1 a
    
15. [pn,ps1]=mapminmax(p);- |3 W7 F5 S3 P
    
16. [tn,ps2]=mapminmax(t);9 H) g, |7 ~\" P) s6 }4 o
    
17. %dx=[-1,1;-1,1;-1,1];/ J! f' n* t: x5 ^5 Y\" e
    
18. [color=Red] p0=minmax(pn);t0=minmax(tn);[/color]; v; w2 ~. x4 s& h7 ?# ~
    
19. - Q9 Q% m) T; s9 Y\" A
    
20. %BP网络训练/ V& z: H2 C* ~; u
    
21. [color=Red] net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); [/color]
    & I, Z, c  T7 B- ]\" _
22. net.trainParam.show=1000;
    8 d! ~' N9 U$ C( O
23. net.trainParam.Lr=0.05;/ J% Q7 r4 M' i: K9 A* q
    
24. net.trainParam.epochs=50000;
    1 k3 J; s* D# Z6 f% o
25. [color=Red] net.trainParam.goal=0.65*10^(-5);[/color]1 _0 h, W  J+ _0 F, l
    
26. net=train(net,pn,tn);
    + u/ j: d5 L9 M; ~5 E8 Z  Y, k
27. 
    0 |0 k2 r* Q2 b1 W7 c7 ?4 E
28. %利用原数据对BP网络仿真
    ! C+ x& O8 q  ^! N4 K\" F
29. an=sim(net,pn);
    6 H8 Y8 A% c0 n9 I8 W
30. a=mapminmax('reverse',an,ps2);
    2 Q) ~8 W& Y3 l1 R2 ^- g: n
31. , m& R, |3 ^. c# y( l, u
    
32. %仿真结果与原数据对比测试 \" k$ F' I1 e' e, f\" N
    
33. x=1990:2009;& I  `2 s5 t\" u
    
34. newk=a(1,:);
    2 s7 f) I, J  A5 p4 C/ [$ k! b
35. newh=a(2,:);
    4 G7 j, D' n9 x\" k4 x, `3 H
36. figure(2);8 |1 ~. u1 ^$ J) c* U
    
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
    + a. U- N2 ^2 m, i
38. legend('网络输出客运量','实际客运量');& Z, T: Y! z* A7 H# r/ a( m' m8 ]
    
39. xlabel('年份');ylabel('货运量、万人');\" w0 F7 H  P, Y' ?. r2 a
    
40. title('运用工具箱客运量学习和测试对比图');
    8 G! s0 F; i4 ?# N; P4 u2 |
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
    1 i$ }+ _, G% G+ @7 B\" |
42. legend('网络输出货运量','实际货运量');
    $ K$ r3 X( H( |3 S6 e6 N! {9 d; A) N
43. xlabel('年份');ylabel('货运量、万吨');
    : n. z. g# c0 A& y\" S
44. title('运用工具箱货运量学习和测试对比图');& Q# K( s; }0 z
    
45. 
      h1 E* g6 U% f
46. %新数据仿真
    , }, W/ \\" z% k! w9 ^4 f$ T% Y
47. pnew=[73.39,75.55# c8 M1 O1 o; Q* h# g/ i( i
    
48.      3.9635,4.09753 D9 ]  B3 B8 V
    
49.      0.9880,1.0268];- ^1 _2 R3 [1 w/ I0 I
    
50. pnewn=mapminmax('apply',pnew,ps1);% O& b- k% A5 _  p) D+ l. T
    
51. anewn=sim(net,pnewn);
    ' }9 M: e& _+ J* F1 N  f4 y! F% I
52. anew=mapminmax('reverse',anewn,ps2)

复制代码

且生

求书名及页码……

T-Eric

gancm 发表于 2014-7-19 12:30
5 C- e/ I6 v7 Q; s( J看了图我感觉是网络输出后面那几个值太大了，前面基本都是一条水平线了，具体哪里还看不出问题。你把修改的 ...

因为版本更新，Matlab中的Newff命令用法有所改变，原命令为：
net=newff(dx,[3,7,2]{'tansig','tansig','purelin'},'traingdx');
我修改为：
0 F! T* [- R! i" q9 R* f2 Cnet=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
然后有些命令被其他命令替代了，其中有premnmx，postmnmx，tramnmx命令好像被替换了，使用了一个强大的命令mapminmax。原命令为：
% V7 |! _" d! ~1 _+ u[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
+ r8 z7 N) V6 W; |a=postmnmx(an,mint,maxt);
pnewn=tramnmx(pnew,minp,maxp);
: h: k# I! ^, uanew=postmnmx(anewn,mint,maxt)
我修改为：
[pn,ps1]=mapminmax(p);[tn,ps2]=mapminmax(t);
" `0 F1 ?0 b6 G0 Aa=mapminmax('reverse',an,ps2);
pnewn=mapminmax('apply',pnew,ps1);
) i# d0 d  c3 Z' Xanew=mapminmax('reverse',anewn,ps2)
( g, m' G% v/ f, W4 x0 N
原程序为：
( Q( v+ m) j- v  _%原始数据输入
, u4 A! r$ r' Q2 U8 }' Aclc
sqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
sqjdcs=[0.6,0.75,0.85,0.90,1.05,1.35,1.45,1.60,1.70,1.85,2.15,2.20,2.25,2.35,...
/ z' o1 I; T/ k& s( n% u3 {* X    2.5,2.6,2.7,2.85,2.95,3.10];
sqglmj=[0.09,0.11,0.11,0.14,0.20,0.23,0.23,0.32,0.32,0.34,0.36,0.36,0.38,0.49,...
  v2 c. K. U: \" p1 x3 E. `- U' ]) _    0.56,0.59,0.59,0.67,0.69,0.79];
" n- n6 }6 ~  ]9 Y; |glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
    22598,25107,33442,36836,40548,42927,43462];
glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
% `) s- h: z5 ?. i  S+ R: d    13320,16762,18673,20724,20803,21804];
p=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];
5 j' n+ C0 }$ d
, r; n% V9 U1 P  {0 L3 a* O%数据归一化
[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
: ]' I* }0 I" V1 S+ Jdx=[-1,1;-1,1;-1,1];

1 W7 V) E: @! l5 i0 H( V; m%BP网络训练
net=newff(dx,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
' E: E; m, ^4 u' q+ t$ anet=train(net,pn,tn);
- E: D/ ?+ ^7 \, b1 v- ]. i
/ R$ H- n4 P. u3 L# [0 S! B%利用原数据对BP网络仿真
- h7 w. J$ C, Nan=sim(net,pn);
a=postmnmx(an,mint,maxt);

%仿真结果与原数据对比测试
x=1990:2009;
) M2 s/ V% x# b3 K; A5 Onewk=a(1,;
" G& ~0 y. x* w, r7 _newh=a(2,;
figure(2);
subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
legend('网络输出客运量','实际客运量');
xlabel('年份');ylabel('货运量、万人');
title('运用工具箱客运量学习和测试对比图');
: R  O% Z, a9 W9 A* E4 _subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
% \) c1 K5 y( h+ z, \legend('网络输出货运量','实际货运量');
xlabel('年份');ylabel('货运量、万吨');
title('运用工具箱货运量学习和测试对比图');
$ C$ s! ?8 d* _1 n+ C8 j, \2 m  K
  z. U9 n/ y% P%新数据仿真
pnew=[73.39,75.55
# g* j, ^5 Y7 N  U9 N    3.9635,4.0975
    0.9880,1.0268];
pnewn=tramnmx(pnew,minp,maxp);
- a; j- w* S  Q3 f: g) I1 xanewn=sim(net,pnewn);
anew=postmnmx(anewn,mint,maxt)
: ]1 V/ S5 b0 W* k9 a/ l
修改后程序为：
%原始数据输入
2 N5 @# |2 ~0 g; i; O  [9 fclc
$ t  F5 r0 `* }. C2 E% [3 W5 Nsqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
sqjdcs=[0.6,0.75,0.85,0.90,1.05,1.35,1.45,1.60,1.70,1.85,2.15,2.20,2.25,2.35,...
* D  c: j, O( }  j( P9 T    2.5,2.6,2.7,2.85,2.95,3.10];
4 D+ A/ q7 r1 I5 H# e* m: S$ X3 Z' Ysqglmj=[0.09,0.11,0.11,0.14,0.20,0.23,0.23,0.32,0.32,0.34,0.36,0.36,0.38,0.49,...
    0.56,0.59,0.59,0.67,0.69,0.79];
4 J9 C6 J5 X" k3 b, Oglkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
    22598,25107,33442,36836,40548,42927,43462];
glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    13320,16762,18673,20724,20803,21804];
' z1 @4 Q4 g. L' o+ p6 I, o5 Sp=[sqrs;sqjdcs;sqglmj];
9 F" Z) Z8 v, K4 q* c% ]- lt=[glkyl;glhyl];

%数据归一化
: H: W3 t5 V- b. |, X; f3 R[pn,ps1]=mapminmax(p);
# V+ x; P# U' `, |7 r[tn,ps2]=mapminmax(t);
- ?" ^# v1 J% X( W" H, e/ j
%BP网络训练
0 B" l# j! H) J1 P- ^  F; a: I8 I4 ^4 k6 Vnet=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
net.trainParam.Lr=0.05;
3 C7 Y4 w* b; P6 Mnet.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
0 S; o4 K: i! `7 D3 Q1 wnet=train(net,pn,tn);
7 x! G# X! z- U3 Z' ~
! Y7 C1 v; y5 W8 c0 F. [. y" e%利用原数据对BP网络仿真
* Z7 t+ K2 h8 C+ qan=sim(net,pn);
" t% X& t+ c2 g4 V# ja=mapminmax('reverse',an,ps2);
1 r8 N9 O4 z5 D
' r: p4 N6 Y* ?5 x%仿真结果与原数据对比测试
: c- N# X* ]. e5 g/ W8 hx=1990:2009;
newk=a(1,;
( N' c+ ?- M' [% Znewh=a(2,;
2 S1 j) d5 d1 F( s, afigure(2);
3 {! J- M- h8 C! qsubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
legend('网络输出客运量','实际客运量');
xlabel('年份');ylabel('货运量、万人');
2 |; X# W0 w5 D! Gtitle('运用工具箱客运量学习和测试对比图');
subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
legend('网络输出货运量','实际货运量');
xlabel('年份');ylabel('货运量、万吨');
title('运用工具箱货运量学习和测试对比图');

%新数据仿真
$ {1 O1 b& x# W' z: M6 spnew=[73.39,75.55
3 W0 y3 x+ V; _5 d    3.9635,4.0975
    0.9880,1.0268];
3 e* `8 u' I- c( ?/ l# Bpnewn=mapminmax('apply',pnew,ps1);
! ]9 i; D0 a3 S' R& X: c" manewn=sim(net,pnewn);
! ?* n& P$ A; Uanew=mapminmax('reverse',anewn,ps2)
（修改的地方用颜色标记了）
麻烦您帮忙指出其中的问题，万分感谢！