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！！

%原始数据输入
clc
, a- U9 E% I7 C0 v( 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,...
2 X; i4 N( j8 T7 g& |4 q6 V    41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
! ~! M  G6 |4 x3 ~% P* H$ 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,...
* Y- P7 ?, P* X# b2 a6 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,...
    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,...
8 W- q0 q+ o& R5 C    22598,25107,33442,36836,40548,42927,43462];
# ?5 a. Y' u. w' N1 w8 H+ tglhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    13320,16762,18673,20724,20803,21804];
p=[sqrs;sqjdcs;sqglmj];
7 R$ A1 b9 t7 it=[glkyl;glhyl];
. x" x8 U/ i7 ?
# o6 D* Q. X6 a8 d+ r+ Q% i%数据归一化
[pn,ps1]=mapminmax(p);
  [$ T4 I% T; A: Z! x  k[tn,ps2]=mapminmax(t);
$ o3 W7 x3 o& wdx=[-1,1;-1,1;-1,1];

%BP网络训练
net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
. m5 @/ r3 X' T2 `+ Q) q+ @% w' Q7 pnet.trainParam.show=1000;
: y, n9 f( N$ ]net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
: G- R! E+ B  X- Y  j5 U' ^net.trainParam.goal=0.65*10^(-3);
net=train(net,pn,tn);
1 X7 f* Y+ `/ H
%利用原数据对BP网络仿真
an=sim(net,pn);
9 o! Z5 U8 ^1 K+ P/ t1 Ia=mapminmax('reverse',an,ps2);

! F, o* \! O: @# `( S% ?  X0 ]9 ^, ^%仿真结果与原数据对比测试
" o& P! N" ]/ @, K' vx=1990:2009;
newk=a(1,:);
newh=a(2,:);
figure(2);
0 }: N) Z2 v5 I- B& V- Q  Asubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
legend('网络输出客运量','实际客运量');
8 x3 z6 B! Q1 o0 _" wxlabel('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--+');
legend('网络输出货运量','实际货运量');
: u5 J( M  }* [1 Uxlabel('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');
/ J# N0 e. ?  j$ o) E
%新数据仿真
pnew=[73.39,75.55
    3.9635,4.0975
# j4 M$ z/ F8 I    0.9880,1.0268];
pnewn=mapminmax('apply',pnew,ps1);
# @( [% e7 c6 \1 k! {& I/ janewn=sim(net,pnewn);
anew=mapminmax('reverse',anewn,ps2)

# f. Z% n4 W* I1 T

T-Eric

狼之魂汪洋 发表于 2014-8-6 14:00
% J/ T8 i' E- I( O) E. Z) f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$ I( q6 `1 S: N: F* y* a
   
2. net=newff(p0,t0,6,{'tansig'},'traingd');
     @. w7 B/ s5 l
3. net.trainParam.show=1000;
   9 r/ \' q# Y4 s) n
4. net.trainParam.Lr=0.05;7 t3 s\" L2 A' t4 ?1 ]# o* F
   
5. net.trainParam.epochs=2000;& B9 q3 Y0 L1 D0 `& P: a
   
6. net.trainParam.goal=0.65*10^(-4);
   * G% E; A\" M& p: ?2 N! |
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
' x5 R- K' r1 s! W三、迭代次数改成2000
/ D. m$ ~0 r7 x上面的参数是自己试的，我也不知道为什么。
$ v+ x, L1 v3 F, ^4 \5 }! e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
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. clc5 ?4 R5 T' u' ?& [7 }- X4 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,...
   % o( B. \\" i7 e& Z' K. @0 \
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];
   9 Q5 V8 u0 a8 i% F) z3 d  z
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,...
   7 G! W+ [2 T) I
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   9 ~+ @; D- l; k9 n% y7 j
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,...
   - T; q  [! v5 a4 N. M* G( y
7.     0.56,0.59,0.59,0.67,0.69,0.79];7 B1 k; U' \- Q6 j- O: n1 t
   
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...+ C* }\" J0 u5 @( @' U
   
9.      22598,25107,33442,36836,40548,42927,43462];5 J1 x% m5 G$ I$ D
   
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...- [: Z: M2 K  n+ h% }
    
11.      13320,16762,18673,20724,20803,21804];9 Y# E; T3 u, @6 Z* n+ V2 t
    
12. p=[sqrs;sqjdcs;sqglmj];* t  ]3 A$ r; n! }
    
13. t=[glkyl;glhyl];$ X* }, c# o, r2 b7 l\" F
    
14. %数据归一化
      d' h0 @5 |$ h* z/ j. p  H
15. [pn,ps1]=mapminmax(p);
    . S\" x; _( `\" v/ ~
16. [tn,ps2]=mapminmax(t);
    2 n' e* d% n8 g' v' l
17. %dx=[-1,1;-1,1;-1,1];
    3 k\" K7 r$ f6 ~: O
18. p0=minmax(pn);t0=minmax(tn);
    . p( _- ^: n# G$ e
19. 
    0 F' T3 N+ C8 o
20. %BP网络训练
    ! Y: V\" J) r& r4 J3 v8 }( R* ]
21. net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); ; L8 S: u( ^9 \! @$ z: n  q
    
22. net.trainParam.show=1000;\" g6 b  M/ P+ a5 u5 A
    
23. net.trainParam.Lr=0.05;4 h' m3 j0 x; C8 q5 i
    
24. net.trainParam.epochs=50000;
    * H9 y; M% @. G
25. net.trainParam.goal=0.65*10^(-5);
    5 ]  J6 x! u8 p) A# a1 [. q
26. net=train(net,pn,tn);
    $ T  Q; J  H; H6 f5 i! H2 l
27. 
    - x& _+ |$ H) V. m& I
28. %利用原数据对BP网络仿真- o# K! ^; n1 `. Y0 V
    
29. an=sim(net,pn);& {5 t  |8 U' {8 t6 N  R* \
    
30. a=mapminmax('reverse',an,ps2);' h, }8 w6 R\" f
    
31. 
    % J) k, q+ n! d\" y* [
32. %仿真结果与原数据对比测试
    - ^. |8 M4 H+ Z0 j5 L8 ?  M
33. x=1990:2009;* b+ S- P9 J' {1 \9 E
    
34. newk=a(1,:);
    - r1 h) K- N! O/ A
35. newh=a(2,:);! U, q' z& s5 M3 u, z
    
36. figure(2);
    : L\" L% `% B' O6 N5 e* D  x
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');' D3 a1 C9 w# ~$ E' T1 ^: U* L, Y
    
38. legend('网络输出客运量','实际客运量');
    $ ?7 w% X* I+ o/ G8 v
39. xlabel('年份');ylabel('货运量、万人');
    ) L2 J\" T) g& d6 r
40. title('运用工具箱客运量学习和测试对比图');% a  U* w: k* A
    
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');5 w1 Y3 v$ y$ U; ?8 \: t# S
    
42. legend('网络输出货运量','实际货运量');4 o- B7 z$ ]! ^3 O5 [5 _
    
43. xlabel('年份');ylabel('货运量、万吨');
    : L3 Q/ N( u) p, y! Y2 X
44. title('运用工具箱货运量学习和测试对比图');
    , J! \/ k+ J+ r. @6 ?+ g; v$ Y
45. 6 e: u1 g: p& I$ I7 u\" U: V; f# I
    
46. %新数据仿真1 _3 Z  q* }. m& g7 C- n
    
47. pnew=[73.39,75.55
    + a% ]& Y7 O5 s; m3 Y3 o1 M6 i
48.      3.9635,4.0975
    ' t* s' t# a) V6 m& T4 ]
49.      0.9880,1.0268];
    7 H- g8 U6 M- Z  t
50. pnewn=mapminmax('apply',pnew,ps1);5 [& A+ g) |4 b1 I
    
51. anewn=sim(net,pnewn);
    5 {8 b* H1 p) c4 B1 m' Y1 P
52. anew=mapminmax('reverse',anewn,ps2)

复制代码

且生

1. clc
   ' X' b! I2 U% Q$ X( G: 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,...
   % v) ~6 E9 w; Y& J  W* R
3.     41.93,44.59,47.30,52.89,55.73,56.76,59.17,60.63];' T: Z) ^1 ^: Q\" l( T9 q
   
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,...
   6 X/ Q$ {% B4 S% D0 N& s9 f
5.     2.5,2.6,2.7,2.85,2.95,3.10];
   , e& _& {\" u  ?\" Z
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,...( P# p8 p$ e' F8 i
   
7.     0.56,0.59,0.59,0.67,0.69,0.79];
   : r4 i! P\" q' D3 f% Y
8. glkyl=[5126,6217,7730,9145,10460,11387,12353,15750,18304,19836,21024,19490,20433,...
     A, U3 y) m8 t' n0 W) C5 ~$ \
9.      22598,25107,33442,36836,40548,42927,43462];
   9 q: L4 p0 c$ Q6 b7 j( ~
10. glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
    \" `. h  u) ~4 J6 |. a2 u4 g
11.      13320,16762,18673,20724,20803,21804];4 m' U7 w$ |6 U: W2 m
    
12. p=[sqrs;sqjdcs;sqglmj];
    $ p8 @0 i& y/ q; p6 [
13. t=[glkyl;glhyl];
    8 I  ^2 ^: n7 p( A! }; ^- X7 n/ g
14. %数据归一化* T9 O' L1 O\" J6 Q6 z  u
    
15. [pn,ps1]=mapminmax(p);/ n+ n7 J% }' @, B' [
    
16. [tn,ps2]=mapminmax(t);9 h2 I0 C( T) x' f9 O/ F- z: V
    
17. %dx=[-1,1;-1,1;-1,1];
    # c\" y. J3 c\" n2 N\" J
18. [color=Red] p0=minmax(pn);t0=minmax(tn);[/color]
    ! S% ^8 X5 F; ^6 R7 C0 Q# a
19. / {5 I6 N3 c! }. X7 T  O2 h
    
20. %BP网络训练
    % L' _2 |8 ^# X
21. [color=Red] net=newff(p0,t0,[3,7,2],{'tansig','tansig','purelin'},'traingdx'); [/color]4 m6 \) \* V* n
    
22. net.trainParam.show=1000;1 ~1 {; [4 U% l\" O6 \$ ^
    
23. net.trainParam.Lr=0.05;' {; A( a7 k. K  ?; Y
    
24. net.trainParam.epochs=50000;, N/ F3 c4 k. e\" r& F, g) @3 {1 A
    
25. [color=Red] net.trainParam.goal=0.65*10^(-5);[/color]
    5 A; W4 [/ U: Y6 |  Z
26. net=train(net,pn,tn);4 a7 }# u! V: r( r- ^
    
27. 
    4 E- o) A( x# }' N9 T' f  K: O( o
28. %利用原数据对BP网络仿真% q7 ~4 l9 _+ r: T' ?. D
    
29. an=sim(net,pn);
    \" \0 A7 J0 f4 ^
30. a=mapminmax('reverse',an,ps2);
    # n& V. N' t& H\" Y
31. 
    ( p+ V9 s: a) W; o/ v; \$ w
32. %仿真结果与原数据对比测试 2 D1 J\" ~- e* K6 v& f* S0 X
    
33. x=1990:2009;
    ; U# i4 }4 S, i9 O8 M4 ]- i
34. newk=a(1,:);* l# {4 A' v4 a# |$ N' c8 A/ y2 l
    
35. newh=a(2,:);
    # Z6 N, |+ t* L- a% o
36. figure(2);2 i' N9 z: @  E) ?
    
37. subplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');2 y; ]/ n% C7 }1 i6 r4 S, r
    
38. legend('网络输出客运量','实际客运量');+ P; O$ V: z  [4 p6 @( H; Q2 A
    
39. xlabel('年份');ylabel('货运量、万人');; P) V+ c% R2 f+ ^9 G
    
40. title('运用工具箱客运量学习和测试对比图');
    ; Z: _& O8 V( _, S5 ?2 [
41. subplot(2,1,2);plot(x,newh,'r-o',x,glhyl,'b--+');
    9 K( k4 p1 F; r0 Z2 s$ [
42. legend('网络输出货运量','实际货运量');
    4 D3 P( W1 W7 Y\" t) @
43. xlabel('年份');ylabel('货运量、万吨');' ?/ K- I/ C: {
    
44. title('运用工具箱货运量学习和测试对比图');
    0 K5 Q& p* X8 p4 U\" |; Y
45. ( e/ w\" S1 t# H
    
46. %新数据仿真: }: N% W$ E1 M. `
    
47. pnew=[73.39,75.55% S2 _- }& p- H9 e
    
48.      3.9635,4.09756 B1 J% l& h# l+ h
    
49.      0.9880,1.0268];
    0 w2 [5 l( ]8 Z7 d) g1 y
50. pnewn=mapminmax('apply',pnew,ps1);3 K3 X! Y+ m( _: h* G9 L! y
    
51. anewn=sim(net,pnewn);; X& |- I+ k9 B8 u
    
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
; G" g) ^1 K, o/ }6 u7 J# R1 y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 ...

" [) D0 ?4 d) ~* X7 G+ a) m因为版本更新，Matlab中的Newff命令用法有所改变，原命令为：
8 e% U+ r, v, \4 \: f" Qnet=newff(dx,[3,7,2]{'tansig','tansig','purelin'},'traingdx');
我修改为：
& R* U  y  g4 v+ ^net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
" {4 @. y- H6 R$ G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：
. O: T  }' @/ L4 b& E7 ]# x[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
a=postmnmx(an,mint,maxt);
pnewn=tramnmx(pnew,minp,maxp);
anew=postmnmx(anewn,mint,maxt)
: q$ v, T5 s. S. A" a我修改为：
( t! ^& H8 }* P0 p8 O[pn,ps1]=mapminmax(p);[tn,ps2]=mapminmax(t);
8 q. S$ D" V# k: U' s2 j# Qa=mapminmax('reverse',an,ps2);
pnewn=mapminmax('apply',pnew,ps1);
anew=mapminmax('reverse',anewn,ps2)
/ ]# j+ A- T3 Z4 z0 @- P( g: {
原程序为：
, k# S8 P. ]% o& z9 J4 `+ J1 W7 _$ s%原始数据输入
clc
4 Q& w/ d/ I. V, U4 ?9 R3 S7 jsqrs=[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 _; k! @9 Y, I4 {2 {; h# O    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,...
    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,...
1 Y) ]& S8 y1 p! v) Z2 s- K    22598,25107,33442,36836,40548,42927,43462];
glhyl=[1237,1379,1385,1399,1663,1714,1834,4322,8132,8936,11099,11203,10524,11115,...
$ q0 i# W: g; @' X+ h4 N# X    13320,16762,18673,20724,20803,21804];
; c4 _/ m7 g3 \" jp=[sqrs;sqjdcs;sqglmj];
t=[glkyl;glhyl];
  k/ @7 v2 H- p8 z* f: T$ ^
& J% F+ n3 a$ ]. V" {. a, |%数据归一化
$ d& u& h: }3 e& n* w[pn,minp,maxp,tn,mint,maxt]=premnmx(p,t);
' {* i" l. e2 Zdx=[-1,1;-1,1;-1,1];

%BP网络训练
' q0 ?; Z. p: q5 ]: ~net=newff(dx,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
1 i7 c2 k  r. l+ f* _net.trainParam.show=1000;
net.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
net.trainParam.goal=0.65*10^(-3);
7 Z6 l' y+ |: w- [* P( v9 ]net=train(net,pn,tn);
( u; x- q- E% Q0 ?& k
%利用原数据对BP网络仿真
' Y6 y) |- Y0 ^$ c3 P4 man=sim(net,pn);
a=postmnmx(an,mint,maxt);

%仿真结果与原数据对比测试
x=1990:2009;
newk=a(1,;
newh=a(2,;
; i1 ^" x/ `$ h. sfigure(2);
3 R/ Q5 n" R2 Z& o( dsubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
! j0 D8 E, f" j7 k( t% alegend('网络输出客运量','实际客运量');
$ B2 z( T. v: t# wxlabel('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 z1 l5 r8 _; e/ Ulegend('网络输出货运量','实际货运量');
% d- ^0 t3 i+ Z2 P( |xlabel('年份');ylabel('货运量、万吨');
; a- D/ J% q5 w# ^) Gtitle('运用工具箱货运量学习和测试对比图');
2 _" ]4 k0 w9 j# O" u
3 _8 @+ a0 B4 C2 E%新数据仿真
pnew=[73.39,75.55
! B. w) \" K) ?8 Q    3.9635,4.0975
    0.9880,1.0268];
6 F, @( b# t1 y1 N8 j. r0 ~pnewn=tramnmx(pnew,minp,maxp);
6 O+ k" A/ q, d4 f9 t! ^1 panewn=sim(net,pnewn);
3 p( @( [. I/ G1 L% n+ ]anew=postmnmx(anewn,mint,maxt)

3 Y* `6 e, U+ ^, p修改后程序为：
%原始数据输入
clc
& W( ^" j% g7 z7 Qsqrs=[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 r9 o+ I  b9 b    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];
5 E5 J: v* O7 |  P% x2 f5 s/ Nsqglmj=[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,...
) e% {! a! O$ s9 s# w/ L    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,...
- g$ ~& g5 A$ N. {% q* e    22598,25107,33442,36836,40548,42927,43462];
) {0 @- a3 L+ {3 C# z3 [8 P" zglhyl=[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  x0 Z7 y3 x5 p' l) at=[glkyl;glhyl];
  `8 U5 ?8 E- o; M9 [) _& e; U0 q
: |/ ?' a) d/ j, `6 p; o* y%数据归一化
1 t% g1 b$ d: S# J2 U; X  F+ S[pn,ps1]=mapminmax(p);
[tn,ps2]=mapminmax(t);

% F" Z- l8 N( n8 G2 y%BP网络训练
net=newff(p,t,[3,7,2],{'tansig','tansig','purelin'},'traingdx');
net.trainParam.show=1000;
6 D9 L9 j; g9 u% U8 ]( X+ d* w0 hnet.trainParam.Lr=0.05;
net.trainParam.epochs=50000;
7 i5 n! v  x6 e7 Znet.trainParam.goal=0.65*10^(-3);
. e0 z" D" A$ Nnet=train(net,pn,tn);

%利用原数据对BP网络仿真
- }$ f$ W$ b9 {( Y9 Z) B6 E: oan=sim(net,pn);
a=mapminmax('reverse',an,ps2);

% w0 T/ N( R' ?* A  ?* [- [%仿真结果与原数据对比测试
x=1990:2009;
newk=a(1,;
) a. V0 r, ?9 [; W: f6 Vnewh=a(2,;
2 ~2 S- F! U# M7 ^/ C- S! J1 [! Rfigure(2);
% M; J' U* ]# h! z* bsubplot(2,1,1);plot(x,newk,'r-o',x,glkyl,'b--+');
legend('网络输出客运量','实际客运量');
/ {$ B# Z& M/ Z( }6 y# Jxlabel('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--+');
" `7 ]5 U: o: P! z3 z- Nlegend('网络输出货运量','实际货运量');
" {: w7 ?% q5 M& Yxlabel('年份');ylabel('货运量、万吨');
0 \9 |" k  A1 k, N: s9 `  |& etitle('运用工具箱货运量学习和测试对比图');
. h3 m& l4 S4 L! x( `
4 X3 B! b) v6 j% l* @( P# r%新数据仿真
pnew=[73.39,75.55
    3.9635,4.0975
    0.9880,1.0268];
# x# ~6 c$ v, H% u- Jpnewn=mapminmax('apply',pnew,ps1);
anewn=sim(net,pnewn);
  n; F6 a1 C: T2 H5 `6 B9 e$ _anew=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！