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！！

3 b& k5 i/ I  u3 T' ?& U%原始数据输入
clc
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,...
1 O$ x) O: H# y9 Z5 W  T; l1 A    2.5,2.6,2.7,2.85,2.95,3.10];
7 T. J+ L: ^( E$ K+ o, V; Ssqglmj=[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];
6 {# U4 t( F, o1 ^glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
    22598,25107,33442,36836,40548,42927,43462];
$ k& ^. c9 o' xglhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    13320,16762,18673,20724,20803,21804];
) v0 a% L" t- ~* t* j' w& L" N3 up=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];

%数据归一化
) g6 M- J  k2 l: ?; A[pn,ps1]=mapminmax(p);
7 l0 x, {  |1 L% Z) b, o' f5 R[tn,ps2]=mapminmax(t);
dx=[-1,1;-1,1;-1,1];

  I9 ~4 y. A- @. V3 z%BP网络训练
/ A9 {+ u  Q4 q8 F4 e* i/ _0 enet=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
' o) K' V' A/ M. d3 S7 vnet.trainParam.Lr=0.05;
, o: z/ }3 o7 xnet.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
! L5 I% k% \7 p' Xnet=train(net,pn,tn);
6 b" O4 G1 l7 b+ p
. ?. D& q4 z7 l%利用原数据对BP网络仿真
an=sim(net,pn);
3 V2 Y2 f0 \! ~7 U$ O4 E( U6 ka=mapminmax('reverse',an,ps2);

%仿真结果与原数据对比测试
% d$ q- V4 V1 {8 ^x=1990:2009;
# Z! }9 e9 O* D* P  M" Z. w. anewk=a(1,:);
newh=a(2,:);
( @6 b" H8 D+ z0 B7 _# i6 }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');
subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
% t( C5 x6 W' o/ r% T/ k6 ?; ~6 Alegend('网络输出货运量','实际货运量');
xlabel('年份');ylabel('货运量、万吨');
, P6 z" f% g' ~: M( A$ V$ ?3 Xtitle('运用工具箱货运量学习和测试对比图');
- ?, y5 @& i* J
%新数据仿真
pnew=[73.39,75.55
    3.9635,4.0975
4 @4 z, {& ?& Y8 f! b    0.9880,1.0268];
: r# J/ ]' z8 U4 P: ?" b7 B; ppnewn=mapminmax('apply',pnew,ps1);
# C6 v. F6 Y; g0 u) o' D3 j. Danewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)

T-Eric

狼之魂汪洋 发表于 2014-8-6 14:00
- u1 G1 c9 g3 v  y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
   6 Z. q: b5 V7 D9 C& |6 Q  Y
2. net=newff(p0,t0,6,{'tansig'},'traingd');
   - x* e  K5 H  I/ l
3. net.trainParam.show=1000;
   \" s8 A9 f6 p. w3 }# }
4. net.trainParam.Lr=0.05;0 v2 G0 n; G3 R# K3 K
   
5. net.trainParam.epochs=2000;
   ; c' p, ]; ]0 y\" |5 f( U8 m- g
6. net.trainParam.goal=0.65*10^(-4);
     r+ s+ p( H. }( s
7. net=train(net,pn,tn);

复制代码

我只改了这里面的，
# P- w8 ]" y9 ~. H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
4 Z+ H+ |. I$ f$ i/ ~* t6 X, P上面的参数是自己试的，我也不知道为什么。
" c* g1 `0 J8 t$ R1 h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。
; ]. e; j& a6 f, k( c' W5 t& r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
% Z! L& x1 Q' O; ?7 M, E忽略我上面那个，改动的地方在18，21，25行……求批

' b3 [& u2 s' i, q; @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. clc8 b$ R7 C9 G% A) w' m
   
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,.../ d4 B' w2 j\" R- ]. X% h6 C
   
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];& T8 J4 l: p$ X* p
   
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 {( r\" O* n0 Y8 N: t* s4 `1 K6 G8 t6 C
5.     2.5,2.6,2.7,2.85,2.95,3.10];3 b& y+ V4 n: n* c, H1 R8 V3 v1 v\" 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,...+ @# Q$ E, C2 u8 n. A: M- H
   
7.     0.56,0.59,0.59,0.67,0.69,0.79];: q' D6 U+ Y2 {4 g5 a
   
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...2 P- N% L5 X- ?# F9 ]0 E2 b
   
9.      22598,25107,33442,36836,40548,42927,43462];
     x. }0 }$ o* _
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    3 d/ _3 z' i- f6 G' c
11.      13320,16762,18673,20724,20803,21804];
    . m9 b- O1 s\" G6 \; }- Q6 a
12. p=[sqrs;sqjdcs;sqglmj];7 c! U  t) C0 A8 A5 v
    
13. t=[glkyl;glhyl];. M% e' K; L1 f# f8 j) @
    
14. %数据归一化1 v$ ~/ {# U+ @4 o; B8 `
    
15. [pn,ps1]=mapminmax(p);; l9 w5 m( P. k
    
16. [tn,ps2]=mapminmax(t);& W5 g* I7 z* T\" T7 z) r7 c5 r2 f
    
17. %dx=[-1,1;-1,1;-1,1];
    . y\" ~9 h: }1 K$ ]
18. p0=minmax(pn);t0=minmax(tn);  }6 j+ v5 T+ b/ V5 l% y
    
19. - s  T  U6 A( |$ s
    
20. %BP网络训练  _4 `3 A* L  s' K5 M
    
21. net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); , t2 P, Z\" O6 X' ?- I! Y
    
22. net.trainParam.show=1000;: f7 X* @! W$ {  K0 o' y5 p
    
23. net.trainParam.Lr=0.05;
    + \# c) s: r% \$ T; K& P
24. net.trainParam.epochs=50000;( d, E4 h6 t- p+ F6 `) Q
    
25. net.trainParam.goal=0.65*10^(-5);5 t- V$ b  }0 K1 p
    
26. net=train(net,pn,tn);9 h- l6 i- A3 ~+ L4 N3 {2 H: o8 n) z/ H
    
27. 
    - }6 B\" ~9 G3 c: \
28. %利用原数据对BP网络仿真% M+ p* A5 d& f3 V+ u' x' ]
    
29. an=sim(net,pn);2 Q2 b! \7 h; c! s6 C' [
    
30. a=mapminmax('reverse',an,ps2);
    8 Q$ |! C, r0 s+ H* _* i
31. 
    - m: G. R4 M% q4 z2 o0 J; z
32. %仿真结果与原数据对比测试 \" r: P! {7 f5 L* s0 Y
    
33. x=1990:2009;: ?/ [' D+ z\" W8 |( C
    
34. newk=a(1,:);& `. ~- i0 k) x- h0 O5 w
    
35. newh=a(2,:);
    # U( z( E+ F/ p  A) q* G
36. figure(2);/ }\" g# M8 S9 m2 Y  u
    
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');0 I; S8 V! L) A
    
38. legend('网络输出客运量','实际客运量');
    1 v\" F& C' Z3 u/ a
39. xlabel('年份');ylabel('货运量、万人');
    $ s7 _+ f7 `) B. O4 E
40. title('运用工具箱客运量学习和测试对比图');% u' R6 w& g& f% X( p, K  j
    
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
    ! F1 i! J! ?; Q1 e, q. V
42. legend('网络输出货运量','实际货运量');
    4 f! b5 |( C% T3 d/ g6 s
43. xlabel('年份');ylabel('货运量、万吨');
    % ]- X9 I/ \5 v# V) R
44. title('运用工具箱货运量学习和测试对比图');\" U: d. C# D; w. `, m
    
45. 
    3 O9 g; K4 V9 m& p
46. %新数据仿真9 a7 t2 h, \9 W* p, w4 r9 P\" g0 K
    
47. pnew=[73.39,75.55
    . V$ Z' r- e5 H: I( p
48.      3.9635,4.0975: [' }! V% w\" w# z
    
49.      0.9880,1.0268];$ ?* \0 I3 F6 W. B, {\" \
    
50. pnewn=mapminmax('apply',pnew,ps1);
    8 X5 Y& V3 [, |, L* g, @. Q3 o
51. anewn=sim(net,pnewn);- R2 W$ o8 P9 g, y' U\" `; p! u+ F6 ]
    
52. anew=mapminmax('reverse',anewn,ps2)

复制代码

且生

1. clc# p3 [8 F( W  G8 {8 A& m& M
   
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,...% n# H\" W9 J8 I0 y- s
   
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];4 j- `4 r3 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,...0 L8 |. `8 `$ J, y& ^\" n+ k0 }* }
   
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   + e7 [/ C7 r3 f. ]
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 `8 H, ?5 K% [+ s6 ]
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   # G# }, e; P# O3 r
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...6 G* _. L1 }% s) V. K2 P) ]  Z\" z
   
9.      22598,25107,33442,36836,40548,42927,43462];% E* n2 }) S1 C\" @  k' w/ W+ n
   
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...5 ~8 |7 f: ^1 I) A% Q, P3 F
    
11.      13320,16762,18673,20724,20803,21804];2 \9 O% j1 P% |' K1 n5 G/ V
    
12. p=[sqrs;sqjdcs;sqglmj];) J  a' g) `' o. A
    
13. t=[glkyl;glhyl];
    + a( O& U1 w0 t+ L8 d
14. %数据归一化4 q, j! y7 m) @. n2 k6 ]
    
15. [pn,ps1]=mapminmax(p);' @) Z4 ~! ?/ o5 t( ]: O2 M+ m
    
16. [tn,ps2]=mapminmax(t);& P- x# I' @! O! ?* [) z
    
17. %dx=[-1,1;-1,1;-1,1];4 [4 j\" M9 m% ^5 x
    
18. [color=Red] p0=minmax(pn);t0=minmax(tn);[/color]
    7 \- g. D* i$ C8 w5 f- d# L& N$ i
19. # z% C\" w& o' ]- r: q  r
    
20. %BP网络训练# d1 _/ U7 E9 I
    
21. [color=Red] net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); [/color]) Z5 W* r4 C: ]/ e9 `
    
22. net.trainParam.show=1000;
    & h# j: Z9 U$ e
23. net.trainParam.Lr=0.05;
    * p& i  f  ?: Y8 \0 q$ H
24. net.trainParam.epochs=50000;9 K# [+ A1 Y) j4 {
    
25. [color=Red] net.trainParam.goal=0.65*10^(-5);[/color]
    0 w. a# V* y# S: k& w4 ^( M
26. net=train(net,pn,tn);' }  u! o7 L6 f' u, v, u4 d
    
27.   u/ K- I\" c6 N4 d: x. i/ `
    
28. %利用原数据对BP网络仿真7 @\" ?# _/ s4 g& \4 @1 c: ~7 M
    
29. an=sim(net,pn);
    9 ~7 i8 n( b. M
30. a=mapminmax('reverse',an,ps2);6 J( a- \% K3 T! ?: s4 _( a: A
    
31. $ v  @\" X' @0 D& x6 ?' }8 j9 l
    
32. %仿真结果与原数据对比测试 ; J% }7 }8 w\" q
    
33. x=1990:2009;5 \+ a, d3 q7 x0 I' S0 r' G
    
34. newk=a(1,:);\" J8 v# O6 t0 _( P3 E
    
35. newh=a(2,:);
    1 _) X% F! I& Y. r\" s- \
36. figure(2);
    5 H! g2 p$ \+ e; E* a
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
    8 O( z1 w/ T6 E8 Z6 F+ H& L
38. legend('网络输出客运量','实际客运量');3 s5 s5 G! R; [2 g) }
    
39. xlabel('年份');ylabel('货运量、万人');
    ; E! C' E1 w  {; x\" k  d
40. title('运用工具箱客运量学习和测试对比图');
    9 H7 S4 \7 C; d$ U: s( j  `8 T: Y
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
    % M. G. p1 C% l\" S6 z6 C, v* q1 f
42. legend('网络输出货运量','实际货运量');
    , t- b1 `9 a0 c8 A
43. xlabel('年份');ylabel('货运量、万吨');
    3 A3 t! [5 b; {0 v; \
44. title('运用工具箱货运量学习和测试对比图');; O2 Q5 k: t$ {/ [) O( h/ K  e
    
45. ; i2 D* _! r- i% ^3 {
    
46. %新数据仿真
    1 Q; [' \, |! f) P5 `7 s5 @6 p( c
47. pnew=[73.39,75.55
    & j7 z4 W9 A! G4 a
48.      3.9635,4.0975
    % u8 \. M8 P+ D7 {\" E' l- q
49.      0.9880,1.0268];
    1 r( ~# E  y( F5 |/ `\" L( H% e
50. pnewn=mapminmax('apply',pnew,ps1);
    5 F3 d2 X: G! i2 d, y4 |+ K
51. anewn=sim(net,pnewn);
    % J( [1 W# l5 `1 ]) I6 J, l
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
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 ...

6 X$ X: q, G9 c# a* P4 `7 H因为版本更新，Matlab中的Newff命令用法有所改变，原命令为：
" U$ O( ^, \- Ynet=newff(dx,[3,7,2]{'tansig','tansig','purelin'},'traingdx');
我修改为：
net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
* i3 o* N3 B* S: v- ~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：
[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
a=postmnmx(an,mint,maxt);
pnewn=tramnmx(pnew,minp,maxp);
" V% m* n4 H) p+ [5 _* panew=postmnmx(anewn,mint,maxt)
我修改为：
[pn,ps1]=mapminmax(p);[tn,ps2]=mapminmax(t);
* T- x  A/ v7 O5 M6 `& Ya=mapminmax('reverse',an,ps2);
$ s  c) h7 A1 D' ^* upnewn=mapminmax('apply',pnew,ps1);
anew=mapminmax('reverse',anewn,ps2)

原程序为：
3 x1 @8 i- i3 j, H/ B%原始数据输入
4 D) b0 @, }  {' @3 [4 Q7 lclc
5 }9 q4 t: i9 W2 H0 d, Usqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
2 R$ I: w, [% u) p    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
2 ~" S/ X8 s! s; H0 Lsqjdcs=[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,...
    2.5,2.6,2.7,2.85,2.95,3.10];
* C$ {- l, K) W1 P5 Asqglmj=[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 m6 N6 R( s% w/ V8 wglkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
. D( l* i5 t8 _, Q5 l: V# y! W    22598,25107,33442,36836,40548,42927,43462];
* P6 M! y& P6 {glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
/ }6 s$ v- M. k5 I1 v' a6 q    13320,16762,18673,20724,20803,21804];
2 U1 R$ @6 Z# Jp=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];

%数据归一化
# B& G8 r* i  h5 |[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
dx=[-1,1;-1,1;-1,1];
8 O% z9 A% e5 S" _
%BP网络训练
net=newff(dx,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
4 B+ d% `/ H6 Q! P4 ]$ pnet.trainParam.show=1000;
net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
- ]8 ^: u/ ~% V& d0 j& g6 Onet.trainParam.goal=0.65*10^(-3);
, r+ R+ u$ T! X, x6 ?) [  Knet=train(net,pn,tn);
' w+ `3 w2 U5 n% |. V
%利用原数据对BP网络仿真
/ i4 O8 b6 z; G; Y  b* |' K& ian=sim(net,pn);
& L; N+ Y+ O7 s* P1 ?a=postmnmx(an,mint,maxt);
# A/ ?, A5 J1 T% b* ^
%仿真结果与原数据对比测试
* w9 E1 ~& B/ R6 k+ i. T4 C1 Y( I6 @x=1990:2009;
newk=a(1,;
newh=a(2,;
0 C0 Y/ Z4 e) ~  p: F. cfigure(2);
9 F# k1 _% Q4 _* K1 L, qsubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
6 W. c; d% t( t8 @1 Rlegend('网络输出客运量','实际客运量');
! f, ~3 h1 w% K- k0 K. Ixlabel('年份');ylabel('货运量、万人');
* ?9 ]- _' j! }: b( btitle('运用工具箱客运量学习和测试对比图');
! F" s5 @( H% A7 r! |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');
# A! a- c/ i* v) mxlabel('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');
+ L* `7 S" M2 \% u0 |6 a
%新数据仿真
) @5 K- S+ ~  w6 s) \) Tpnew=[73.39,75.55
* E  B9 d0 G" N3 w7 b" i    3.9635,4.0975
    0.9880,1.0268];
5 C. Q4 R& D# ^& L( j* Spnewn=tramnmx(pnew,minp,maxp);
, n5 ?) C8 n* Panewn=sim(net,pnewn);
anew=postmnmx(anewn,mint,maxt)
7 H3 F9 z8 _4 f
修改后程序为：
%原始数据输入
clc
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,...
, o# V7 D6 v: }    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,...
    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,...
3 T2 J& q9 Y4 {9 M- K    0.56,0.59,0.59,0.67,0.69,0.79];
! ?/ J9 h' t% l* j7 e  }% gglkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
; @: t. K$ i) o' _8 m% g/ U5 C    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];
p=[sqrs;sqjdcs;sqglmj];
; h! m( [. G$ r& T: Dt=[glkyl;glhyl];
) v7 H7 i" c3 O, n
; |0 g, o- F; o% L%数据归一化
[pn,ps1]=mapminmax(p);
[tn,ps2]=mapminmax(t);

9 {% V, p: C% c$ U( \& C- P3 S: N0 [%BP网络训练
net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
2 o. O5 M% G# I+ _1 ^8 g4 Anet.trainParam.show=1000;
net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
* y( C7 S  n6 v4 lnet.trainParam.goal=0.65*10^(-3);
net=train(net,pn,tn);
" |$ K" I( ]6 z3 x) W& c" O+ c
- A5 u1 w* f- d' E& I9 p6 j% G%利用原数据对BP网络仿真
& N4 D7 M% Q* X$ U& K* jan=sim(net,pn);
a=mapminmax('reverse',an,ps2);
) p" Y0 ]. u: |% g5 Z; g$ p4 P' a
, i" F8 x" J8 Q$ @- T/ v1 ]- S- j%仿真结果与原数据对比测试
x=1990:2009;
1 n) S. v/ }! cnewk=a(1,;
newh=a(2,;
4 c+ b1 s5 `" u: r) B4 t: Ufigure(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');
, h. D' d; ~& Ttitle('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');
9 O9 K3 v( a, q5 f: {title('运用工具箱货运量学习和测试对比图');

; n7 S6 j; m& o1 i' ^* X1 T1 I%新数据仿真
pnew=[73.39,75.55
    3.9635,4.0975
    0.9880,1.0268];
pnewn=mapminmax('apply',pnew,ps1);
anewn=sim(net,pnewn);
! V7 W$ X$ f. q  k* @# U4 janew=mapminmax('reverse',anewn,ps2)
（修改的地方用颜色标记了）
$ Q3 G/ _5 X  i* i, l* d% I麻烦您帮忙指出其中的问题，万分感谢！