bpnet2.m

来自「Matlab写得神经网络的训练实例」· M 代码 · 共 19 行

function F=BPnet2()
X=-1:0.1:1;T=sin(pi*X);
plot(X,T,'+');

[R,N]=size(X);[S2,N]=size(T);S1=5;
[W1,b1,W2,b2]=initff(X,S1,'tansig',S2,'purelin');
[y1,y21]=simuff(X,W1,b1,'tansig',W2,b2,'purelin');

disp_freq=10;
max_epoch=8000;
err_goal=0.02;
lr=0.01;
tp=[disp_freq max_epoch err_goal lr];
[W1,b1,W2,b2,te,tr]=trainbp(W1,b1,'tansig',W2,b2,'purelin',X,T,tp);
[y1,y22]=simuff(X,W1,b1,'tansig',W2,b2,'purelin');
plot(X,T,'-',X,y21,'--',X,y22,'o');
X1=0.5;
[y1 y2]=simuff(X1,W1,b1,'tansig',W2,b2,'purelin');
y2