自编神经收集法度模范.docVIP

tic p0=[0,0,0,0,0.007,0.0091,0.0096,0.0123,0.0143,0.0183,0.0254,0.0312,0.0379,0.0475,0.0441,0.0452,0.056,0.0808,0.1307,0.1704,0.2002,0.2291,0.2494,0.2841; 0,0,0,0.007,0.0091,0.0096,0.0123,0.0143,0.0183,0.0254,0.0312,0.0379,0.0475,0.0441,0.0452,0.056,0.0808,0.1307,0.1704,0.2002,0.2291,0.2494,0.2841,0.2985; 0,0,0.007,0.0091,0.0096,0.0123,0.0143,0.0183,0.0254,0.0312,0.0379,0.0475,0.0441,0.0452,0.056,0.0808,0.1307,0.1704,0.2002,0.2291,0.2494,0.2841,0.2985,0.3292; 0,0.007,0.0091,0.0096,0.0123,0.0143,0.0183,0.0254,0.0312,0.0379,0.0475,0.0441,0.0452,0.056,0.0808,0.1307,0.1704,0.2002,0.2291,0.2494,0.2841,0.2985,0.3292,1.0987]; t=[0.007,0.0091,0.0096,0.0123,0.0143,0.0183,0.0254,0.0312,0.0379,0.0475,0.0441,0.0452,0.056,0.0808,0.1307,0.1704,0.2002,0.2291,0.2494,0.2841,0.2985,0.3292,0.3721,1.0987]; b={[tansig],[logsig],[purelin]}; c={[trainlm],[trainbfg],[traingd],[learngdm]}; d={[learngd],[learngdm]}; for s=5:24 for i=1:length(b) for n=1:length(b) for j=1:length(c) for m=1:length(d) net=newff(minmax(p0),[s,1],{b{i},b{n}},c{j},d{m},mse); %训练参数设置，可供选择参数为： %传递函数：tansig\logsig\purelin, %反向传播网络的训练函数：trainlm\trainbfg\traingd\learngdm, %权重与偏量反向传播的学习函数：learngd\learngdm, %显示函数：mse\mseeg [net,tr]=train(net,p0,t); p1=[0.3292,0.3721,0.435,0.5557; 0.3292,0.3721,0.435,0.7048; 0.3721,0.435,0.5557,0.8877; 0.435,0.5557,0.8877,1.0987]; a=sim(net,p1); a1=[0.5557,0.7048,0.8877,1.0987]; o=(abs(a(1)-a1(1))/a1(1)+abs(a(2)-a1(2))/a1(2))/2; if o=1e-02 disp(网络训练正常) break else disp(网络训练异常) end end if o1e-02 disp(网络训练异常，请重新设置训练参数) end end end end end p2=[0.5557; 0.7048; 0.8877; 1.0987]; disp(我国固定资产投资额为：) gte=sim(net,p2) toc