1、0.1000-0.39300.8000-0.31200.66000.2000-0.16470.9000-0.2189-0.40000.46090.3000-0.09881.0000-0.3201例一、利用三层 BP 神经网络来完成非线性函数的逼近任务,其中隐层神经元个数为五个。样本数据:解:看到期望输出的范围是(-1,1),所以利用双极性 Sigmoid 函数作为转移函数。程序如下: clear; clc;X=-1:0.1:1;D=-0.9602 -0.5770 -0.0729 0.3771 0.6405 0.6600 0.4609.0.1336 -0.2013 -0.4344 -0.5000
2、 -0.3930 -0.1647 -.0988.0.3072 0.3960 0.3449 0.1816 -0.312 -0.2189 -0.3201;figure;plot(X,D,*); %绘制原始数据分布图(附录:1-1) net = newff(-1 1,5 1,tansig,); net.trainParam.epochs = 100; %训练的最大次数net.trainParam.goal = 0.005; %全局最小误差net = train(net,X,D); O = sim(net,X); figure;,X,O); %绘制训练后得到的结果和误差曲线(附录:1-2、1-3)V
3、= net.iw1,1%输入层到中间层权值theta1 = net.b1%中间层各神经元阈值W = net.lw2,1%中间层到输出层权值theta2 = net.b2%输出层各神经元阈值所得结果如下:输入层到中间层的权值:中间层各神经元的阈值: 中间层到输出层的权值:V = (-9.1669 7.3448 7.3761 4.8966 3.5409)Tq= (6.5885 -2.4019 -0.9962 1.5303 3.2731)TW = (0.3427 0.2135 0.2981 -0.8840 1.9134)输出层各神经元的阈值: T = -1.5271例二、利用三层 BP 神经网络来完
4、成非线性函数的逼近任务,其中隐层神经元个数为五个。48215396107看到期望输出的范围超出(-1,1),所以输出层神经元利用线性函数作为转移函数。X = 0 1 2 3 4 5 6 7 8 9 10;D = 0 1 2 3 4 3 2 1 2 3 4;2-1) net = newff(0 10,5 1,purelin) net.trainParam.epochs = 100; net.trainParam.goal=0.005;net=train(net,X,D); O=sim(net,X);2-2、2-3)V = (0.8584 2.0890 -1.2166 0.2752 -0.3910
5、)T-9.83407.4331-2.01350.5610)T-1.12342.32084.6402-2.2686)q= (-14.0302中间层到输出层的权值:W = (-0.4675 T = 1.7623例三、以下是上证指数 2009 年 2 月 2 日到 3 月 27 日的收盘价格,构建一个三层 BP 神经网络,利用该组信号的 6 个过去值预测信号的将来值。日期价格2009/02/022011.6822009/03/022093.4522009/02/032060.8122009/03/032071.4322009/02/042107.7512009/03/042198.1122009/0
6、2/052098.0212009/03/052221.0822009/02/062181.2412009/03/062193.0122009/02/092224.7112009/03/092118.7522009/02/102265.1612009/03/102158.5722009/02/112260.8222009/03/112139.0212009/02/122248.0922009/03/122133.8812009/02/132320.7922009/03/132128.8512009/02/162389.3922009/03/162153.2912009/02/172319.442
7、2009/03/172218.3312009/02/182209.8622009/03/182223.7312009/02/192227.1322009/03/192265.7612009/02/202261.4822009/03/202281.0912009/02/232305.7822009/03/232325.4812009/02/242200.6522009/03/242338.4212009/02/252206.5722009/03/252291.5512009/02/262121.2522009/03/262361.7012009/02/272082.8522009/03/2723
8、74.44clear;D1=2011.682 2060.812 2107.751 2098.021 2181.241 2224.711.2319.442.2206.572.2221.082.2128.851.2325.481.2374.44;D = premnmx(D1)%数据归一化把数据化到-1,1范围内Q=length(D);count = 1:1:Q; X=zeros(6,0); X(1,2:Q)=D(1,1:(Q-1);X(2,3:(Q-2);X(3,4:(Q-3);X(4,5:(Q-4);X(5,6:(Q-5);X(6,7:(Q-6);plot(count,D,count,D,3-1
9、) net = newff(minmax(X),7 1,plot(count,D,count,O,r3-2、3-3) -2.4916-3.0098 0.4381 1.1598 0.6343-0.3355 -0.08990.68491.29690.27820.9312-0.2707-1.72260.94451.7617-0.21672.11460.9514-0.83880.22140.1251-0.1086-0.7422-0.1918-0.43111.58000.65191.97480.2787-0.78190.72380.0084-0.7738-2.1268-1.0499-2.17401.2349 V = 1.2791 -0.6276 0.4532 -0.0884 -1.6249q= (2.6717 1.9258 -0.0286 -1.2134 -1.0657 0.8908 1.6032)TW = (2.66281.0361 -1.41602.2844 -0.3706 -1.4939 -1.4575) T = -0.5480
copyright@ 2008-2022 冰豆网网站版权所有
经营许可证编号:鄂ICP备2022015515号-1