小波神经网络预测的代码1

((net(j)-b(j))/b(j))/a(j);%计算结束Р%调整d_a(j)РendРendР%step4--------网络重新计算-------------------------------------------РWij=Wij-lr*d_Wij;РWjk=Wjk-lr*d_Wjk;Рb=b-lr*d_b;Рa=a-lr*d_a;Р%修正各个权值Рy=zeros(1,N);%输出节点初始化Рnet=zeros(1,n);%隐形节点初始化Рnet_ab=zeros(1,n);%隐形节点初始化Рfor i=1:1:NРfor j=1:1:nРfor k=1:1:MРnet(j)=net(j)+Wjk(j,k)*x(k);Рnet_ab(j)=(net(j)-b(j))/a(j);РendРy(i)=y(i)+Wij(i,j)*_ab(j));Р%mymorlet是judyever编写的小波函数,以后可以扩展成输入不同的小波名字即可Р% y(i)=mysigmoid(2,y(i));РendРendРepoch=epoch+1;Рerr=d-y;РSSE=err*err'/M;Р[ epoch SSE]РendР%step5--------输出-------------------------------------------Рplot(x,d,'r',x,y,':'); Рtitle('训练后的目标序列和实际输出序列');Р% gtext({'This is the first line','This is the second line'})Р% gtext({'First line','Second line'},'FontName','Times','Fontsize',12)Рlegend('target output','WNN output',1);