%% 使用优化后的权值和阈值测试结果

%% 使用优化后的权值和阈值

inputnum=size(P,1);%输入层神经元个数

outputnum=size(T,1);%输出层神经元个数

N=size(P_test,2);

M=size(P,2);

%% 新建BP

net=newff(P,T,9);

%% 设置网络参数：训练次数1000，训练目标0.001，学习速率00.1

net.trainParam.epochs =3000;

net.trainParam.goal = 1e-6;

net.trainParam.lr = 0.01;

%% BP初始权值和阈值

w1num=inputnum*hiddennum;%输入层到隐含层的权值个数

w2num=outputnum*hiddennum;%隐含层到输入层的权值个数

w1=bestX(1:w1num);%初始输入层到隐含层的权值

B1=bestX(w1num+1:w1num+hiddennum);

w2=bestX(w1num+hiddennum+1:w1num+hiddennum+w2num);%初始隐含层到输出层的权值

B2=bestX(w1num+hiddennum+w2num+1:w1num+hiddennum+w2num+outputnum);%输出层阈值

net.iw{1,1}=reshape(w1,hiddennum,inputnum);

net.lw{2,1}=reshape(w2,outputnum,hiddennum);

net.b{1}=reshape(B1,hiddennum,1);

net.b{2}=reshape(B2,outputnum,1);

%% 训练网络

net=train(net,P,T);

%% 测试网络

t_sim_P= sim(net,P);

t_sim_P_test= sim(net,P_test);

%% 反归一化

T=mapminmax('reverse',T,ps_output);

T_sim_P= mapminmax('reverse',t_sim_P,ps_output);

T_sim_P_test = mapminmax('reverse',t_sim_P_test,ps_output);

%% 相对误差

error_P=abs(T_sim_P-T)./T;

error_P_test=abs(T_sim_P_test-T_test)./T_test;

%% 相关系数

R2_P= (M * sum(T_sim_P .* T) - sum(T_sim_P) * sum(T))^2 / ((M * sum((T_sim_P).^2) - (sum(T_sim_P))^2) * (M * sum((T).^2) - (sum(T))^2));

R2_P_test = (N * sum(T_sim_P_test .* T_test) - sum(T_sim_P_test) * sum(T_test))^2 / ((N * sum((T_sim_P_test).^2) - (sum(T_sim_P_test))^2) * (N * sum((T_test).^2) - (sum(T_test))^2));

%% 结果对比

result = [T_test' T_sim_P_test' abs(T_test-T_sim_P_test)']

result=[T' T_sim_P' abs(T-T_sim_P)']