UntitBP24.zip_load forecasting_电力 负荷预测_电力负荷曲线_神经电力负荷_负荷曲线MATLAB

clear p=[4.567,5.7369,6.1161;5.7369,6.1161,9.2105;6.1161,9.2105,13.8258;9.2105,13.8258,19.5787;13.8258,19.5787,17.0507;19.5787,17.0507,16.998;17.0507,16.998,17.7732;16.998,17.7732,20.7335;17.7732,20.7335,26.268;20.7335,26.268,33.1576;26.268,33.1576,29.43;33.1576,29.43,28.0715;29.43,28.0715,34.0092;28.0715,34.0092,45.898;34.0092,45.898,59.0953;45.898,59.0953,68.1366;59.0953,68.1336,78.1489;68.1336,78.1489,72.688;78.1489,72.688,84.7979;72.688,84.7979,96.0597;84.7979,96.0597,107.3389;96.0597,107.3389,121.8486;107.3389,121.8486,139.9338;121.8486,139.9338,163.5153]'; t=[9.2105,13.8258,19.5787,17.0507,16.998,17.7732,20.7335,26.268,33.1576,29.43,28.0715,34.0092,45.898,59.0953,68.1366,78.1789,72.688,84.7979,96.0597,107.3389,121.8486,139.9338,163.5153,173.0294]; [pn,ps]=mapminmax(p); [tn,ts]=mapminmax(t); net=newff(pn,tn,10); net.trainParam.epochs=3000; %训练次数 net.trainParam.goal=0.001; %训练所要达到的精度 net.trainParam.lr=0.1; %学习速率 net=train(net,pn,tn); an=sim(net,pn); a=mapminmax('reverse',an,ts); for i=1:24 e(i)=t(i)-a(i); c(i)=e(i)/t(i); end d=mean(abs(c)); a e c d plot(1:24,t,'b:',1:24,a,'r-') legend('实际值', '预测值') xlabel('年份/年') ylabel('用电量/亿KW*h') title('BP神经网络负荷预测曲线') figure plot(c,'b-*') xlabel('年份/年') ylabel('负荷误差') title('BP神经网络负荷预测误差变化图')