bp神经网络数字识别

function bpgdtrain global bpnet; clc; load templet pattern; c=0; p=[]; for i=1:10 for j=1:20 c=c+1; p(:,c)=pattern(i).feature(:,j); end end