Satya Narayan Shukla

Could you please post the training logs here?

I'm looking into the logs and checking the code, would be updating the code shortly to get the 0.858 ROC.

No, there is no need to standardize the data. The different scales of different dimension are accounted for In the autoencoder loss.

yeah, the format seems right. Make sure the mask variable is 1 where x is observed else 0, the starting value of T should be 0 (basically subtract the initial...

You can try to run the data extraction for each variable in a parallel way. Basically, every ``curr.execute`` command in the code can be executed parallelly.

Hi @yvonneywu and @IcurasLW, please take a look at the MIMIC-III dataset page and it has instructions to create the database.