Results closeloop and doubts about this modality

1 次查看（过去 30 天）

显示更早的评论

FRANCISCO 2013-2-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/65265-results-closeloop-and-doubts-about-this-modality

关闭： MATLAB Answer Bot 2021-8-20

在 MATLAB Online 中打开

Well I've done some testing with closeloop, though at first with all data overfitting caused me, I have reduced the size of the data and obtained the following results in openloop:

MSE=6,28470e-05  
MSEa=0,000121796
R2=0,974646308527986
R2a=0,951088280194053  
R2trn=0,982303160187683  
R2val=0,890073733738057  
R2tst=0,923433417603953

These values I have obtained with the following code:

if true
  % code
 MSE = perform(net,targets,outputs);
MSEa=Neq*MSE/(Neq-Nw);
R2=1-MSE/MSE00;
R2a=1-MSEa/MSE00a;
MSEtrn=tr.perf(end);
MSEval=tr.vperf(end);
MSEtst=tr.tperf(end);
R2trn=1-MSEtrn/MSE00;
R2val=1-MSEval/MSE00;
R2tst=1-MSEtst/MSE00;
end

well as the result with closeloop is next, and I think quite disastrous for the accuracy achieved in openloop network:

Well I tried to upload a picture but I have not been able. Is a graph in which 10 iterations compared with real targets and not really look anything. To which may be due?. I have used 250 data and I do not overtighten. I would like to calculate the precision indicarais as in closeloop and if anyone knows if I explain closeloop really disastrous results are obtained. thank you very much