search

addb-swstarlab / DATATune

DATATune (Databse Parameter Tuning via Autoendoer Latent Space)
11 stars 3 forks source link

sample scaler question #4

Closed HwanheeKim813 closed 4 months ago

HwanheeKim813 commented 5 months ago

Hello,

In the code, the scale of newX seems to be different from the scale of the data on which the regressor was trained. So I wonder why you use the new scaler Z_scaler instead of X_scaler. Is there another reason I'm not noticing?

Thank you.

Kwon-sein commented 4 months ago

Thank you for your interest in my code.

The reason why the scaler of newX is different from the scaler of the regressor (I understand that TabNet) is that the input data of the regressor is original data X, but new_X data is the augmented data. It means two input data are different That's why they didn't use the same scaler.

If you have any additional questions, please leave an issue.

thank you.