the routine MLloop expects and a normalized(or?) RMSE to be returned from the routine evaluation_map. Evaluation_map doesn't compute nor returns a the nRSME, thus the Tool Task3 crashes.
I propose to add the computation of a normalized RMSE on Line 39, like this: nRMSE=np.sqrt(mean_squared_error(comp_Y[:,0],comp_Y[:,1])/(comp_Y[:,1].max()-comp_Y[:,1].min()))
Question 1: Is that the correct way to compute the relative RMSE in the routine?
Question 2: what is referred to in the code as relative MSE is the normalized RMSE - correct?
dsgoll123
commented
1 year ago
Hi
the routine MLloop expects and a normalized(or?) RMSE to be returned from the routine evaluation_map. Evaluation_map doesn't compute nor returns a the nRSME, thus the Tool Task3 crashes.
I propose to add the computation of a normalized RMSE on Line 39, like this:
nRMSE=np.sqrt(mean_squared_error(comp_Y[:,0],comp_Y[:,1])/(comp_Y[:,1].max()-comp_Y[:,1].min()))Question 1: Is that the correct way to compute the relative RMSE in the routine?
Question 2: what is referred to in the code as relative MSE is the normalized RMSE - correct?