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！！
# `1 E, }5 w6 d) {, u% D: F
%原始数据输入
clc
+ _- c$ h, q: k: T/ c  U# {* L+ m/ isqrs=[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 O9 c% W# a+ F. M$ ?  b7 a" G5 L3 Y    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
, @* l" W9 i) tsqjdcs=[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,...
4 D/ L/ `# i# i) }    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,...
    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];
! y+ G& V; T" {2 N& aglhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    13320,16762,18673,20724,20803,21804];
p=[sqrs;sqjdcs;sqglmj];
9 i2 m6 T1 M3 y  C' [t=[glkyl;glhyl];
% y& q0 Z$ t1 Y0 G! i# G7 r* F
  L8 c8 @' m$ ]7 V%数据归一化
$ l; j& f. v& Z' [, d& j[pn,ps1]=mapminmax(p);
. Z- {7 m3 N& T0 C8 L/ ~[tn,ps2]=mapminmax(t);
dx=[-1,1;-1,1;-1,1];

/ c9 \9 W  `; Q( L%BP网络训练
: R' M/ K' d1 T, R, d. mnet=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
! X+ f7 L4 C3 K# P% n# ^net.trainParam.show=1000;
" {0 ]; ^$ h0 f3 _) t, ]4 R3 Qnet.trainParam.Lr=0.05;
% ~/ K2 b  v4 H1 G+ o* Bnet.trainParam.epochs=50000;
. t4 V) J9 M2 W$ k7 j  ?$ {net.trainParam.goal=0.65*10^(-3);
net=train(net,pn,tn);
- S1 Z8 U, I) z- ?1 K' @
" |3 U" I* m1 c4 X1 t  D$ @# a. Q4 b%利用原数据对BP网络仿真
9 s0 b6 K& t% J* V& {an=sim(net,pn);
. h! U/ E' w! |' j( U+ F; m1 [/ z; Xa=mapminmax('reverse',an,ps2);
" ~: b: j$ a# l  N, n4 V  ~# c
' Z+ k' L+ |3 s: X+ l( ?%仿真结果与原数据对比测试
' ~+ F8 p5 o0 `* ^x=1990:2009;
newk=a(1,:);
newh=a(2,:);
figure(2);
) }. M! s. S! M; s, wsubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
legend('网络输出客运量','实际客运量');
, a. f4 b6 t$ r+ Cxlabel('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--+');
- J0 l6 d, `- U6 t: t& V6 ylegend('网络输出货运量','实际货运量');
2 S! o3 I. Z; j( z/ Yxlabel('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');

* B/ b7 D! a% ?/ S  d- h+ N% K%新数据仿真
+ D9 V  P! |* a) Lpnew=[73.39,75.55
    3.9635,4.0975
    0.9880,1.0268];
pnewn=mapminmax('apply',pnew,ps1);
anewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)

/ |% i9 o& P1 F$ |7 @' x7 y0 e

T-Eric

狼之魂汪洋 发表于 2014-8-6 14:00
! Z! n8 F  F& m4 _5 s8 J4 e你这例题是在什么书上找的？

7 N/ ^! w, g5 W3 Z2 t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
   \" ~; N: j6 K6 ~. c; w' a
2. net=newff(p0,t0,6,{'tansig'},'traingd'); ; D$ e  Y  F: b: O# e5 p\" m
   
3. net.trainParam.show=1000;+ Z! u$ N2 p) X) O8 `7 N. N
   
4. net.trainParam.Lr=0.05;
   ( e' U' t6 s2 j3 ^# C\" I& b\" S  E
5. net.trainParam.epochs=2000;/ ~\" j! m. \  @! y
   
6. net.trainParam.goal=0.65*10^(-4);
   8 l+ E7 M$ ?1 G; `4 J8 k
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
3 [$ `- ~' J# z- u+ x0 z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。
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。
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
8 @" @% U5 T% E. D8 |0 y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
   % f& ?4 b( F# O# `5 N
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* ~4 V1 u( o0 w
   
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];4 B' U0 X9 S+ n* K\" }
   
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,...\" o# {/ F' `! o9 a% z% Z+ l+ S
   
5.     2.5,2.6,2.7,2.85,2.95,3.10];/ Q4 B8 ^3 Q- N/ {& a$ h$ 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 i- [( ], S+ J0 {/ _% E$ o
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   6 D6 k! G' K$ m# z4 h, H/ n
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...0 G\" f, p6 i- T/ Q- ]2 U+ e* S
   
9.      22598,25107,33442,36836,40548,42927,43462];
   ! `( ?' U# q9 I+ B  t- c
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...5 V$ I3 M- c7 P\" ?  q
    
11.      13320,16762,18673,20724,20803,21804];3 P. s\" l7 `1 o2 b/ x
    
12. p=[sqrs;sqjdcs;sqglmj];
    ( V, l0 B+ r  v: z; L  [% f
13. t=[glkyl;glhyl];0 D- y/ U1 H. z$ W
    
14. %数据归一化- s( K& c/ |3 r. H2 s\" m\" }
    
15. [pn,ps1]=mapminmax(p);8 M; R8 r$ P5 W$ H
    
16. [tn,ps2]=mapminmax(t);0 T, [$ q% j' {) d
    
17. %dx=[-1,1;-1,1;-1,1];5 l8 h/ W) w6 c. k8 ?
    
18. p0=minmax(pn);t0=minmax(tn);
    6 l! m* R2 w3 V6 G( m  ]# o
19. 
    9 [\" A3 t! h; J& i; z
20. %BP网络训练
    + I- N2 a/ \# K5 y% e\" b! n
21. net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); ( W$ [/ E+ p- p9 `/ W
    
22. net.trainParam.show=1000;8 x; H2 T$ @' n
    
23. net.trainParam.Lr=0.05;; x% H8 [$ t; x( M1 C1 Q9 K. ]8 I$ z
    
24. net.trainParam.epochs=50000;
    5 m# ^# L- q; R- R' z1 t) [3 u  n# }
25. net.trainParam.goal=0.65*10^(-5);
    8 y0 S& }+ G( ?\" U
26. net=train(net,pn,tn);, R  y+ I1 ]  V! e. X
    
27. & J4 m3 h$ c& a/ V  H. b' E
    
28. %利用原数据对BP网络仿真) _0 V/ |/ O0 q4 n8 I5 i0 d
    
29. an=sim(net,pn);2 E9 m& `  ^8 w# o$ p
    
30. a=mapminmax('reverse',an,ps2);) z3 E  K) a' R# E# Y- E5 w
    
31. ( ]% K; [# u' I
    
32. %仿真结果与原数据对比测试
    ; V$ I! K: {; |& R6 {
33. x=1990:2009;
    : G' b* X- k* T' y! b
34. newk=a(1,:);. O5 a- c3 n: w5 `
    
35. newh=a(2,:);
    5 m\" Z  }+ q* B! v; T, }
36. figure(2);
    6 H0 L4 X) B- r5 O3 u2 d
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
    \" V4 W# ]- m2 r: k+ X
38. legend('网络输出客运量','实际客运量');
    - ^( @5 H1 I* U( C6 T1 V
39. xlabel('年份');ylabel('货运量、万人');
    ; H; b: u\" M7 Z2 Y
40. title('运用工具箱客运量学习和测试对比图');
    ) W1 Q1 p; c) I# s) h
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');% f7 w\" L; g8 a( y0 h# n
    
42. legend('网络输出货运量','实际货运量');
      n) o' N# q1 G; y7 E! N
43. xlabel('年份');ylabel('货运量、万吨');
    4 K\" }! A* E: F( L; ^+ |9 c- C
44. title('运用工具箱货运量学习和测试对比图');
    2 y0 A9 C: j2 u
45. 
    . b4 O% |0 R& A) Q8 H
46. %新数据仿真: f5 p# a5 E$ W2 N
    
47. pnew=[73.39,75.55( V2 W7 H\" h- u7 k3 i+ G8 J
    
48.      3.9635,4.09750 u4 v2 L. z% g% F/ Y' T
    
49.      0.9880,1.0268];$ x+ b9 H0 h+ a' V\" H# D
    
50. pnewn=mapminmax('apply',pnew,ps1);
    / h5 m( T9 o% N
51. anewn=sim(net,pnewn);
    + ]4 p, Y6 X, m8 P% K$ r
52. anew=mapminmax('reverse',anewn,ps2)

复制代码

且生

1. clc; ]5 ]- `+ E; L/ J# {9 x& f+ W+ 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,...
   , F% S& Q' j6 |# p: H, n, U, m9 R2 E
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
   ( \7 l$ F' G7 L9 z9 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,...& G, N* b% ~2 w\" I
   
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   ( |6 _& U9 w# Z, 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,...
   + e4 A) H7 H; W6 t\" S1 x9 @, t% R
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   2 `% j8 A8 l1 `2 g! B$ b
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...4 ?! ]* \0 k- ^5 d! x6 k8 ^7 Y
   
9.      22598,25107,33442,36836,40548,42927,43462];
   7 E6 h! X0 Y' x5 t: v9 \
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    ) F; D, x. ?1 \. J
11.      13320,16762,18673,20724,20803,21804];
    4 e; \8 m9 {' W6 L$ g! b0 }0 U
12. p=[sqrs;sqjdcs;sqglmj];
    3 f9 m; A( \\" h' I
13. t=[glkyl;glhyl];
    : y- b0 H  F# {( b; _& m
14. %数据归一化$ F$ \& L4 v5 Q( {\" b/ T
    
15. [pn,ps1]=mapminmax(p);, b$ |; o1 r( {3 Q7 K( ~' j
    
16. [tn,ps2]=mapminmax(t);4 [; F% n. F4 }9 h8 }\" w7 d5 Q
    
17. %dx=[-1,1;-1,1;-1,1];, ?\" V6 w! \+ l5 Q
    
18. [color=Red] p0=minmax(pn);t0=minmax(tn);[/color]
    4 _' F* O: n# c7 m! M6 e1 }
19. 
    8 s; J! D7 S6 s
20. %BP网络训练3 P/ V* i, |% q% J: w
    
21. [color=Red] net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); [/color]
    , V6 Y& d4 ^1 T3 m& |# X
22. net.trainParam.show=1000;\" `4 D2 {3 D1 u
    
23. net.trainParam.Lr=0.05;6 C# L1 a\" A1 Y
    
24. net.trainParam.epochs=50000;# C6 m3 |1 W* ]
    
25. [color=Red] net.trainParam.goal=0.65*10^(-5);[/color]
    / p9 ^3 p  q( C, ?7 ]9 _$ |0 ~, ]
26. net=train(net,pn,tn);5 B1 _8 g* `5 C; t
    
27. 
    # N) Y9 _6 i; ^7 H# F1 w& f+ ?
28. %利用原数据对BP网络仿真% B& `0 L3 h/ O8 t- W
    
29. an=sim(net,pn);
    ; E% F9 ^5 Q: l5 h% R% @
30. a=mapminmax('reverse',an,ps2);
    9 W4 R4 h2 Q/ f\" e- d
31. + R3 d% \/ x5 j
    
32. %仿真结果与原数据对比测试 6 [, X7 Q$ [) W4 E
    
33. x=1990:2009;
    ; N' y0 M( E2 |, Z% ^\" L  t( @! a
34. newk=a(1,:);
    - S8 m: f2 ?- |/ Z$ S& k
35. newh=a(2,:);; @- S( `/ Q; a( p8 D\" e
    
36. figure(2);- S: W) E3 [2 X  x
    
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');$ c( o1 n3 k3 a: H
    
38. legend('网络输出客运量','实际客运量');\" D* _: P' A: R7 |2 g9 R/ O  [
    
39. xlabel('年份');ylabel('货运量、万人');; V* Y/ ?9 J' h; y& v7 O  l2 q
    
40. title('运用工具箱客运量学习和测试对比图');
    : l. P8 ~5 ]0 M* ?
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');2 {3 {* K( i! f3 r3 H0 j
    
42. legend('网络输出货运量','实际货运量');4 e# ^3 I. [, G2 J
    
43. xlabel('年份');ylabel('货运量、万吨');
    5 @: @# L* ]/ T8 f2 I
44. title('运用工具箱货运量学习和测试对比图');: a/ G* Z* d- d
    
45. : P- d% D1 F, t6 |0 X4 V\" ]- }
    
46. %新数据仿真8 }6 `1 w3 @4 k' l
    
47. pnew=[73.39,75.55% b/ ~: }) |4 u) i% _
    
48.      3.9635,4.0975% N) E8 l8 C( i; r# i
    
49.      0.9880,1.0268];
    . t9 ]+ e' f1 C
50. pnewn=mapminmax('apply',pnew,ps1);; r* x9 C/ ~. \! o; u
    
51. anewn=sim(net,pnewn);- }# Y: h0 ^; k+ ]
    
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
' D% \7 b( Q* ?, Y* e3 _% [9 ?& M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：
, x; f# g* H. D5 ]3 e9 B$ a9 X- Bnet=newff(dx,[3,7,2]{'tansig','tansig','purelin'},'traingdx');
我修改为：
# `% T6 M1 s; B, v" T( @; Unet=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：
& N6 N- M8 M( P8 ]3 j[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
% p) ]% e7 D% b& ~4 S! ^% F$ m8 E, V" Fa=postmnmx(an,mint,maxt);
pnewn=tramnmx(pnew,minp,maxp);
+ I9 M- F' d) w: I, ranew=postmnmx(anewn,mint,maxt)
我修改为：
  |  L- ~3 }1 B4 R3 ]; b0 J[pn,ps1]=mapminmax(p);[tn,ps2]=mapminmax(t);
: J* k* E: S7 Y9 Aa=mapminmax('reverse',an,ps2);
pnewn=mapminmax('apply',pnew,ps1);
anew=mapminmax('reverse',anewn,ps2)

原程序为：
8 x) o" O4 [5 X5 [%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,...
0 i, o& r' S( a4 ]    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
0 d0 c4 |6 U; r( usqjdcs=[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,...
    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,...
" c' T9 B5 `* p- r" Y    22598,25107,33442,36836,40548,42927,43462];
2 D, N1 ]7 B$ f  g0 W. b" {; y" W1 hglhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
6 h, n: _- ~: k' j5 A    13320,16762,18673,20724,20803,21804];
3 k# H% D9 i8 Sp=[sqrs;sqjdcs;sqglmj];
' G0 n. e+ q, ~1 V. ]t=[glkyl;glhyl];
- S3 g) l" o( Q$ [
%数据归一化
1 L% X: {6 l: U' R$ x9 ^[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
dx=[-1,1;-1,1;-1,1];

8 ?! H  g  z( l% l: ~* _2 W%BP网络训练
4 D% |( \' f* c7 G! V% g3 unet=newff(dx,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
: ]5 @8 I) W0 k6 q4 \, }7 F+ Unet.trainParam.goal=0.65*10^(-3);
, s4 k4 O( j! d. L6 S! Gnet=train(net,pn,tn);
4 J8 [# L) X; ^' G! `2 u
, ^# @9 O! a! v/ W  t%利用原数据对BP网络仿真
3 c2 T: A! c5 @+ T& l' gan=sim(net,pn);
) B1 s/ S# m. D; ra=postmnmx(an,mint,maxt);
* M1 ]% @3 l: ?( i, t1 M4 @; o
%仿真结果与原数据对比测试
x=1990:2009;
7 K! N( I% [3 Y, V7 Xnewk=a(1,;
newh=a(2,;
figure(2);
subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
. O+ Y. _' e8 X9 @legend('网络输出客运量','实际客运量');
8 I: K% N7 e; a/ xxlabel('年份');ylabel('货运量、万人');
& i8 @1 m" _$ ^% j8 utitle('运用工具箱客运量学习和测试对比图');
% S2 M9 ]. n7 \% W$ l$ H$ _* ?subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
+ L- u# i. X3 qlegend('网络输出货运量','实际货运量');
  \5 ~" e% |* B! G0 K4 B. \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');

8 H  \. I+ }+ B* v%新数据仿真
pnew=[73.39,75.55
# r' N# l# ?, e3 g( s* t    3.9635,4.0975
    0.9880,1.0268];
pnewn=tramnmx(pnew,minp,maxp);
anewn=sim(net,pnewn);
anew=postmnmx(anewn,mint,maxt)

3 [' l6 P) I& u4 e6 w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,...
% V' t6 C# f) N    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
% L5 a% ^' z) w# R3 v# U3 ^# N7 Osqjdcs=[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];
) u5 P9 p# ^: d: A/ t1 a" ?) 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,...
  e4 N' F! V5 o$ S    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];
! ^# K4 J' k0 p, u4 e9 _% {! h# Nglhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
6 L) |  |" h' F2 D3 H    13320,16762,18673,20724,20803,21804];
p=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];
+ ^! D  e) n% j! d& i' X: ]& E
3 p8 I8 r* I& T' B4 M, g. A# r0 C%数据归一化
[pn,ps1]=mapminmax(p);
! g3 \5 O) L" S* }" ], e; M[tn,ps2]=mapminmax(t);
# Y9 R- K% F/ ]! J+ X+ X
%BP网络训练
net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
0 F# k- j9 H. S2 Fnet.trainParam.show=1000;
net.trainParam.Lr=0.05;
6 Z0 b5 _# O2 Snet.trainParam.epochs=50000;
# G9 O+ M* s, U& S  Y& W1 ~* h0 Cnet.trainParam.goal=0.65*10^(-3);
, H. H2 U1 |" _: ]2 q' {net=train(net,pn,tn);
$ \+ h4 g% d  X
%利用原数据对BP网络仿真
an=sim(net,pn);
a=mapminmax('reverse',an,ps2);
2 ]8 C% j9 b5 }$ m5 k0 n
0 G2 o- E8 Z8 @' V* N& \7 d7 g+ s5 Q%仿真结果与原数据对比测试
x=1990:2009;
newk=a(1,;
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');
subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
" h/ R! d2 q0 v* P; q- C1 Zlegend('网络输出货运量','实际货运量');
. t% O( Q& \. dxlabel('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
pnew=[73.39,75.55
+ }, Y+ F8 Y4 l. X0 q, _    3.9635,4.0975
    0.9880,1.0268];
+ y+ z8 }% K% @' Tpnewn=mapminmax('apply',pnew,ps1);
anewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)
: {# G. k1 C2 e* L6 X$ a' T（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！