Evaluating neural network prediction formulas

Matthew Clark

2019 3 23

0 个回答

7 次查看（30 天）

0 个投票

Hi there I need some help,, I am evaluating my neural network prediction with MAPE,MAE,MSE,RMSE methods are these formulas below correct ? What values I should use to evaluate performance NORMALIZED or DENORMALIZED data ?

r^2 is 0.94, MSE is very high. Suggestions ?

%where DENORM(obtained DENORMALIZED data after NN prediction, elLoad-original data(but not normalized))
%for MAE
err=denorm-elLoad;
ae=abs(e);
maeErr=mean(ae);
%for MAPE
mapeErr=mean(abs(e/elLoad));
%for MSE
mseErr=mean(denorm-elLoad).^2;
%for RMSE
mseErr=sqrt(mean(denorm-elLoad).^2);