求帮忙

lyfsgdhr

P=[0.1093,0.1110,0.1127,0.1141,0.1154,0.1164,0.1171,0.1175,0.1178,0.1179,0.1179,0.1179,0.1179,0.1180,0.1182];
T=[0.1110,0.1127,0.1141,0.1154,0.1164,0.1171,0.1175,0.1178,0.1179,0.1179,0.1179,0.1179,0.1180,0.1182,0.1185];
6 K/ X4 _3 ?: W; A8 i%  创建一个新的前向神经网络
net_1=newff(minmax(P),[10,1],{'tansig','purelin'},'traingdm')
0 c# v3 C' E* c! y$ P7 g! Y%  当前输入层权值和阈值
( Q1 T' q! p# J5 D8 E1 GinputWeights=net_1.IW{1,1}
inputbias=net_1.b{1}
%  当前网络层权值和阈值
layerWeights=net_1.LW{2,1}
( e# w% E$ E. t, q$ D, xlayerbias=net_1.b{2}
8 \( s7 }& O; t+ t- m%  设置训练参数
net_1.trainParam.show = 50;
$ l) ~4 O- L9 @2 ]! i* l2 K+ ?! enet_1.trainParam.lr = 0.05;
net_1.trainParam.mc = 0.9;
! s- d& Q% h- C4 E' I; T5 Wnet_1.trainParam.epochs = 10000;
# s! u( B5 Y: R  Qnet_1.trainParam.goal = 1e-3;
! ^( U" h4 n% w9 W/ E%  调用 TRAINGDM 算法训练 BP 网络
[net_1,tr]=train(net_1,P,T);
- F1 l* w6 v9 m  G% d/ A%  对 BP 网络进行仿真
A = sim(net_1,P);
  B' C4 f0 `6 x$ D. ~- M%  计算仿真误差
$ Q/ q2 s. K" P% `$ cE = T - A;
3 v$ S! m5 }0 Y$ QMSE=mse(E)
x=[0.1110,0.1127,0.1141,0.1154,0.1164,0.1171,0.1175,0.1178,0.1179,0.1179,0.1179,0.1179,0.1180,0.1182,0.1185]';%测试
sim(net_1,x)
这段程序是根据14年的数据，来预测下一年的，怎么算不出来啊 。

fffffffddf

这个太难了，我也不会

346010696

眼早就花了~~~

renzhe1

强烈支持。楼主万岁

rchjia

哦～～

tzw

楼主的帖子实在是写得太好了。可是我立刻想到，这么好的帖子，倘若别人看不到，那么不是浪费楼主的心血吗？经过痛苦的思想斗争，我终于下定决心，牺牲小我，奉献大我。我要拿出这帖子奉献给世人赏阅，我要把这个帖子一直往上顶，往上顶！顶到所有人都看到为止！  

xuwei198215

哦～～

heyingfeifei

楼主的帖子实在是写得太好了。可是我立刻想到，这么好的帖子，倘若别人看不到，那么不是浪费楼主的心血吗？经过痛苦的思想斗争，我终于下定决心，牺牲小我，奉献大我。我要拿出这帖子奉献给世人赏阅，我要把这个帖子一直往上顶，往上顶！顶到所有人都看到为止！  

yaotl17

留个脚印```````

lihao54

来报道!!!!!!!!!!!