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！！
# {8 ]! N) U! {9 a- x
0 B) M; m$ p7 `( I9 \7 H%原始数据输入
" t7 P( v5 K' U0 I2 f$ Fclc
  \/ S0 q/ _# c4 k' N6 U. D& ?0 {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,...
0 q" a0 j3 _1 p3 U# G, ]    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,...
( h9 V8 k! M9 e1 s8 l    2.5,2.6,2.7,2.85,2.95,3.10];
, s6 c. t5 U. H0 V: r1 x- }$ G# Xsqglmj=[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];
- b  }+ E# [  A' o$ Lglkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
1 I3 i- h; ^0 `3 F8 P4 e    22598,25107,33442,36836,40548,42927,43462];
% H$ a9 b& P) X( Oglhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
( M  m6 a2 l% b    13320,16762,18673,20724,20803,21804];
$ T1 ]( {9 u0 q0 v( W/ }( kp=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];
; ]/ a/ k$ _' H# c% O2 k
%数据归一化
[pn,ps1]=mapminmax(p);
[tn,ps2]=mapminmax(t);
dx=[-1,1;-1,1;-1,1];

%BP网络训练
net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
6 `/ S$ d" M2 m. Unet.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
net=train(net,pn,tn);

2 L% {% m2 D& r' ~%利用原数据对BP网络仿真
an=sim(net,pn);
1 w8 c9 g( m% Y- N. i+ X1 Sa=mapminmax('reverse',an,ps2);

! [& Y( t3 _' Z%仿真结果与原数据对比测试
x=1990:2009;
: S3 t5 M- h# T' V/ I) m, I( Nnewk=a(1,:);
newh=a(2,:);
( W# h2 E8 S8 R# Yfigure(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');
1 ~, x$ G8 D' b( N5 Ltitle('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');
0 U" H( E  x9 z0 `" wxlabel('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');
  T9 h8 s. I& _
6 f" n7 k2 x6 Z* I7 H* X%新数据仿真
1 Z0 P; s1 e% J. y' f, h& A( fpnew=[73.39,75.55
    3.9635,4.0975
2 ?+ \' t. }8 m8 E- _, S    0.9880,1.0268];
pnewn=mapminmax('apply',pnew,ps1);
+ z" Y# A% O3 X- N' {! ]( H! yanewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)
8 X6 h' K) I  r8 Q( v7 r" E& m

T-Eric

狼之魂汪洋 发表于 2014-8-6 14:00
! k. K2 v3 g0 c0 {你这例题是在什么书上找的？

. \8 w. U6 ^  K! `) s1 k$ }7 Z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( z6 A. @; Q* _/ W
   
2. net=newff(p0,t0,6,{'tansig'},'traingd');
   0 ^' A/ N5 g* ^/ h
3. net.trainParam.show=1000;
   * I# @9 q8 t5 R- J1 }
4. net.trainParam.Lr=0.05;
   2 V# h- B4 S0 E$ j0 h
5. net.trainParam.epochs=2000;
   5 u! r# I: u) j8 c* c\" r
6. net.trainParam.goal=0.65*10^(-4);4 G( B) @) t$ n7 l
   
7. net=train(net,pn,tn);

复制代码

我只改了这里面的，
; ^# e( c' E& q" x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
  I( m" F7 N( m' h( Y! v上面的参数是自己试的，我也不知道为什么。
0 q3 f0 H: o9 b; D2 s0 \! M: _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。
- ~2 u! `1 J5 I* s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
4 O( p" h; g! E) w( `: |9 B1 R忽略我上面那个，改动的地方在18，21，25行……求批

- A' g' q! }* _/ I" K, {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
   ' p3 I  {( V2 v' m\" P5 }2 \0 e+ \! |
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,...
   ( \7 a; `7 |, H1 {7 J
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];- q5 f) J2 B  N; I& N3 \3 I
   
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 z; g% n1 p  R7 ~
   
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   : e\" y. n+ u# G+ e9 P4 t
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,...2 M* e# O) B& Y0 X$ ?
   
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   3 D% a, i( L+ Z- U  w& y7 a
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...# U; w- G* n' h, c2 r5 y
   
9.      22598,25107,33442,36836,40548,42927,43462];
   . N/ t* t0 X) I4 M/ z: u
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    6 r/ J6 |: Y: w3 s2 Q\" T
11.      13320,16762,18673,20724,20803,21804];
    . }8 I* l# ]5 G3 E
12. p=[sqrs;sqjdcs;sqglmj];! _) O% h$ u& D: p
    
13. t=[glkyl;glhyl];' @! d; t+ }+ [' U
    
14. %数据归一化( x5 l0 ^. E) d- E1 a4 Y! S% E\" v& K
    
15. [pn,ps1]=mapminmax(p);  l0 r# K: p; R+ X% V- w
    
16. [tn,ps2]=mapminmax(t);
    2 K5 _5 n% k* Q  u, l
17. %dx=[-1,1;-1,1;-1,1];
    \" A# t0 \9 w% m2 ^2 V
18. p0=minmax(pn);t0=minmax(tn);
    % z7 D% \8 A8 J
19. 
    . g. I6 S% n8 w* H6 M) G
20. %BP网络训练
    4 v+ K+ r) J6 x
21. net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); 3 c* D# I8 r3 G  a/ `
    
22. net.trainParam.show=1000;. ?7 [% I' y4 F; U# Z
    
23. net.trainParam.Lr=0.05;
    5 B) f1 ?5 f1 Q3 z8 e8 Z
24. net.trainParam.epochs=50000;; H/ C8 J: d* S# m\" {* i
    
25. net.trainParam.goal=0.65*10^(-5);
    9 f6 w' A% _& k( G
26. net=train(net,pn,tn);. C4 E% H- o. L4 j
    
27. : y9 d' n* H, {6 T2 d5 X- V$ V2 x
    
28. %利用原数据对BP网络仿真- R. C1 ^: W0 U& {/ B/ K$ a
    
29. an=sim(net,pn);
    1 Q\" l3 b$ o, M% T
30. a=mapminmax('reverse',an,ps2);8 N4 D2 k+ \- t8 ]+ M3 S
    
31. 
    3 |! ~: x$ V8 e' [' {
32. %仿真结果与原数据对比测试
    8 a+ V3 F3 r7 N6 @- ?6 e
33. x=1990:2009;
    8 J+ i$ q+ C. {/ _
34. newk=a(1,:);
    2 F/ N5 O, u5 o% _
35. newh=a(2,:);
      p7 B' z/ o0 d* d$ F2 S$ ~
36. figure(2);; o! u/ j4 P\" {1 U) }$ |
    
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');2 k) v+ o$ i2 Z. t4 Y% x
    
38. legend('网络输出客运量','实际客运量');5 I5 z) p& [+ O) d% R
    
39. xlabel('年份');ylabel('货运量、万人');3 d# O- @! b4 m; F' }
    
40. title('运用工具箱客运量学习和测试对比图');7 O- F, G) o  Y+ |) C
    
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');* d: m$ s7 Q\" E
    
42. legend('网络输出货运量','实际货运量');, Q3 c2 k- G, k, J4 I
    
43. xlabel('年份');ylabel('货运量、万吨');\" t3 C\" {8 U: Q
    
44. title('运用工具箱货运量学习和测试对比图');0 O. ?+ D( G1 d) o* {3 N
    
45. 
    4 |. N4 ?% V; J) [0 }' {1 ?9 D
46. %新数据仿真! q4 N1 s; d6 ^( y, F
    
47. pnew=[73.39,75.55+ E! {. O$ K$ m, ^5 v* u\" ^
    
48.      3.9635,4.0975% K! P$ R; I- N& K0 U
    
49.      0.9880,1.0268];
    9 A% M) D! b1 N$ E) v# b
50. pnewn=mapminmax('apply',pnew,ps1);: g1 V6 Q3 k, [5 J
    
51. anewn=sim(net,pnewn);
    . H4 Q# s0 L3 N1 y( e5 q
52. anew=mapminmax('reverse',anewn,ps2)

复制代码

且生

1. clc
   ' }: ?7 h  h4 \& b\" S5 y
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,...
   ' Y0 ?1 T. Y- a. u
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
   % s0 N2 e\" \' F; J1 P, W7 h% i7 r
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,...
   ) }/ d* s( {5 y5 T# R  P2 m
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   : e2 `+ j' X8 [\" o7 {: ~; X
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,...
   ! m\" t) o0 q/ q9 S6 u7 n: J! K
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   + y9 D- F1 G' X/ t) Y( h* M5 s
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...\" r* B2 R, I$ c: `% u) @
   
9.      22598,25107,33442,36836,40548,42927,43462];) g' |! S& j' C& w1 X/ n
   
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    : J# e! E8 j6 E( @
11.      13320,16762,18673,20724,20803,21804];. f# `0 z. O% e\" J  r8 R
    
12. p=[sqrs;sqjdcs;sqglmj];8 H6 T- |$ y\" E
    
13. t=[glkyl;glhyl];\" ^. Z& U3 G2 U- K. x+ _
    
14. %数据归一化0 Q% ]; ~: Q$ Y( S' C
    
15. [pn,ps1]=mapminmax(p);
    7 w3 J( r2 |9 i0 V9 y
16. [tn,ps2]=mapminmax(t);0 V( t$ {3 p( x( ]) X
    
17. %dx=[-1,1;-1,1;-1,1];! h8 q; I1 }. {3 l; `# J1 F\" z  e
    
18. [color=Red] p0=minmax(pn);t0=minmax(tn);[/color]
    4 l, z' a# ^2 J3 m
19. . K3 O* e0 b+ W1 T% {$ \4 m4 {) n6 G% _3 A
    
20. %BP网络训练9 X8 L% q: E, y+ L% {- C- ^, u
    
21. [color=Red] net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); [/color]
    - F& Y' B0 K& ~8 A5 p
22. net.trainParam.show=1000;/ S: U7 \; a  ^) g
    
23. net.trainParam.Lr=0.05;\" E3 x& b5 a2 e% @
    
24. net.trainParam.epochs=50000;) u0 }4 I* |8 ^  m0 {0 G& @
    
25. [color=Red] net.trainParam.goal=0.65*10^(-5);[/color], y6 O* l  e6 Z1 q9 `
    
26. net=train(net,pn,tn);
    - b8 ]9 D% Z) w
27. : o2 r+ {7 |* ?4 s/ w6 R+ k\" s# C
    
28. %利用原数据对BP网络仿真
    : R* I( N1 f, Q  X7 X  c: t
29. an=sim(net,pn);* Y6 I3 c- Z3 h/ y- T; H3 r0 O
    
30. a=mapminmax('reverse',an,ps2);
    ( a: d* b+ T& q: N1 w
31. : ]2 J- W) r7 V' [\" W
    
32. %仿真结果与原数据对比测试
    7 x: d9 p* q/ D8 y$ z
33. x=1990:2009;5 O- L# @+ \& l$ R. ^, R0 ]3 t
    
34. newk=a(1,:);) n, Z( o+ N7 E/ b
    
35. newh=a(2,:);
    # |* M3 V) J* t+ x( j: f- Y
36. figure(2);4 Q- P7 r9 q\" U. V6 u, b
    
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
    4 }1 t/ y+ k: s
38. legend('网络输出客运量','实际客运量');
    , o8 T2 f# v8 a! F: d3 C
39. xlabel('年份');ylabel('货运量、万人');
    5 m# d) [8 I- `* e. l* B8 x
40. title('运用工具箱客运量学习和测试对比图');
    + `1 L$ }' {, f: F. K
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');9 ^% K* w0 k+ h3 S) h2 \% N
    
42. legend('网络输出货运量','实际货运量');
    ; U* A4 T1 U6 n5 v6 f
43. xlabel('年份');ylabel('货运量、万吨');# s( A2 E9 Y# X: O, H/ R3 O  z
    
44. title('运用工具箱货运量学习和测试对比图');
    0 g4 I# r! O- {; D
45. 
    ) I6 H0 W7 y7 r
46. %新数据仿真
    7 w/ {\" t, p2 O9 v
47. pnew=[73.39,75.55
    ' l/ N+ I8 ?3 Y; e* z
48.      3.9635,4.0975- v/ a/ J$ E) m  u) I, d
    
49.      0.9880,1.0268];. a2 D$ N0 K# O- ?. y; t1 H, |
    
50. pnewn=mapminmax('apply',pnew,ps1);
    ; v; e: }( q( z6 c! [9 I
51. anewn=sim(net,pnewn);! i; N: S- I% Z; E4 K- g# S& o
    
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 ...

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：
% L9 Z* R8 H- h* r8 t8 J* Znet=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：
# b& h6 Q; Q+ u4 [& F+ u. \[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
; ?* [; l% k8 w, @4 f7 ya=postmnmx(an,mint,maxt);
4 y) Q9 B3 p/ d/ U! Lpnewn=tramnmx(pnew,minp,maxp);
2 c1 _2 N4 q% V2 [  o0 A+ A; u7 oanew=postmnmx(anewn,mint,maxt)
+ j9 H# @! l. g/ m) V9 J我修改为：
" ~2 J. l+ j% E! P2 ]5 h[pn,ps1]=mapminmax(p);[tn,ps2]=mapminmax(t);
a=mapminmax('reverse',an,ps2);
  s3 G; Y- D. l" J& Upnewn=mapminmax('apply',pnew,ps1);
anew=mapminmax('reverse',anewn,ps2)

原程序为：
% t- ~+ h9 p+ ~5 I+ ?2 q; P%原始数据输入
& q) Q" x' V( ?7 R8 e' zclc
6 g/ ~* b3 O/ ?8 j! P8 vsqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
) g0 j5 u5 h, C* [6 O7 q. d5 s    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,...
* V/ r! [, m, Q& }3 R0 F    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,...
6 L0 x0 y+ x, J    0.56,0.59,0.59,0.67,0.69,0.79];
/ h5 R6 k+ M. s% e. c% Kglkyl=[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];
p=[sqrs;sqjdcs;sqglmj];
; x, }( f/ `5 M  qt=[glkyl;glhyl];

%数据归一化
[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
3 ~1 \; G' g% ]: [dx=[-1,1;-1,1;-1,1];
. |' @/ _! O, U4 _: I6 u6 X9 u
%BP网络训练
" k$ p2 J8 I% u2 Z# Q/ W5 v. Ynet=newff(dx,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
$ J. u/ ]; a7 v0 u) Mnet.trainParam.show=1000;
& _6 r+ K0 r/ U% C: }net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
$ ~3 g4 V, Q( ~net.trainParam.goal=0.65*10^(-3);
( t  x2 X2 [; X- W) V3 O+ A3 D2 Tnet=train(net,pn,tn);
& G( q1 i. j# e% f5 J0 G: o3 q
) x% i3 [9 ?! }4 |" m" m%利用原数据对BP网络仿真
  O- |% S+ \1 X8 ?: U" l+ san=sim(net,pn);
a=postmnmx(an,mint,maxt);

%仿真结果与原数据对比测试
2 r! `0 L9 u5 V" G( H$ wx=1990:2009;
newk=a(1,;
, T( o8 |- e  b! K# M2 Q+ Hnewh=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');
* _, U4 U. U7 a* F7 o. |5 Uxlabel('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');
! A, A& R: x: Z, P/ `subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
) [! k8 ], \, u& z8 ]# |7 T- ylegend('网络输出货运量','实际货运量');
- q3 }. o! C0 X( Nxlabel('年份');ylabel('货运量、万吨');
( S) K  u! P$ {title('运用工具箱货运量学习和测试对比图');
3 p$ N# ]' o1 I. V) C
%新数据仿真
pnew=[73.39,75.55
$ C+ S! Y% P3 W7 F    3.9635,4.0975
    0.9880,1.0268];
0 ~2 }% @& @) b. B' O; f# v5 lpnewn=tramnmx(pnew,minp,maxp);
5 w# |& d" [" }/ a) G/ ^! v4 C2 x5 Kanewn=sim(net,pnewn);
4 ?/ j7 K: m& z/ O: s3 Y6 Q' Banew=postmnmx(anewn,mint,maxt)

5 e. N& ^, H( @  a) z) o修改后程序为：
%原始数据输入
clc
, F$ f/ T$ e/ Asqrs=[20.55,22.44,25.73,27.13,29.45,30.1,30.96,34.06,36.42,38.09,39.13,39.99,...
' ^1 [$ b8 k, p0 v' A    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,...
$ ]/ S# T. a$ {7 T, n    2.5,2.6,2.7,2.85,2.95,3.10];
* f! H0 W2 d* b% e" E# usqglmj=[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];
glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
    22598,25107,33442,36836,40548,42927,43462];
/ G8 J( s8 j( u- U, c3 X9 c3 _glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
0 a6 \6 g) L. R0 c% v, B2 m    13320,16762,18673,20724,20803,21804];
4 E$ q& x% a6 r% d: ^p=[sqrs;sqjdcs;sqglmj];
% ^. e8 Q. p% R9 P0 It=[glkyl;glhyl];
3 ^9 S. @8 X' d% W9 _
%数据归一化
[pn,ps1]=mapminmax(p);
[tn,ps2]=mapminmax(t);
3 |/ c) z( _! I; o7 Q" {, W
& ~( S6 j/ `( ~%BP网络训练
5 B' X: Y7 Q$ q" C( enet=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
net.trainParam.Lr=0.05;
: u6 N, p- x* `& z/ g# \net.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
" D5 c6 p: g/ I6 S' x) Xnet=train(net,pn,tn);

- o) B# `9 C! \% X%利用原数据对BP网络仿真
an=sim(net,pn);
a=mapminmax('reverse',an,ps2);

4 M4 J: h7 W2 o  R%仿真结果与原数据对比测试
x=1990:2009;
* F: l' ~+ O" g6 Tnewk=a(1,;
2 o- j2 C5 Z# C/ G& f0 o" Unewh=a(2,;
1 U+ Z- ?" V6 l3 D+ ~# Q7 s* |5 Dfigure(2);
2 V$ p7 m2 o1 J1 l% R7 N- fsubplot(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');
* l7 M' N5 X. a3 O6 K% f+ D5 }3 M8 Rtitle('运用工具箱客运量学习和测试对比图');
& l  ^; b" t; u6 \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');
" x5 x6 K+ ]& X3 J' Cxlabel('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');

  }/ }$ }* ?3 ~%新数据仿真
pnew=[73.39,75.55
    3.9635,4.0975
    0.9880,1.0268];
pnewn=mapminmax('apply',pnew,ps1);
9 {# o& G( X1 Sanewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)
（修改的地方用颜色标记了）
; b4 d1 Y# b8 u5 a5 c麻烦您帮忙指出其中的问题，万分感谢！