edebrouwer / gru_ode_bayes Goto Github PK

Hi there!

The following files are referenced in the MIMIC preprocessing README
but are not present in the repo:

• mimic_preproc.py
• folds_split_mimic.py

Can you please point me to their location/can you add them to the repo?

Best, Dan

As mentioned in several prior issues (#1, #2, #3 and #4) the final MIMIC preprocessing files mimic_preproc.py and folds_split_mimic.py are not included in the data_preproc/MIMIC/ folder.

It'd be great to gain complete insight into how you set up the EHR data with this project. The text in Appendix K of your paper only goes so far.

Thanks!

while running double_OU from experiments/double_OU folder I get the following error

fileNotFoundError: [Errno 2] No such file or directory: '../../gru_ode_bayes/datasets/double_OU/double_OU_metadata.npy'

When running climate.gruode.py, the error occurs.
p_obs = p[i_obs] IndexError: tensors used as indices must be long, byte or bool tensors

How to solve this problem?

According to paperswithcode.com this seems to be the state of the art at time series forecasting (in general).
Is it the state of the art at forecasting of financial stock forecasting too?
If so, would it gain more money than it would loose in average?

If it would still not be lucrative, what value of MSE and NEGLL would be needed to ensure on average lucrativity (winning money)?

I'm stuck at this stage while experimenting with the model, I'd appreciate it if you could help.

The referenced files folds_split_mimic.py and mimic_prepoc.py are missing from the repo

Hello,

I have no legal background, but according to currently according to this discussion: https://opensource.stackexchange.com/questions/1720/what-can-i-assume-if-a-publicly-published-project-has-no-license,

If a repository has no license, then all rights are reserved and it is not Open Source or Free. You cannot modify or redistribute this code without explicit permission from the copyright holder.

It would be great to add an appropriate LICENSE to this repository so that others can re-use the code, if you want to allow that, or to clarify explicitly if you want to reserve all rights.

Hi I got this error when I am trying to 'import grud_ode_bayes' .

Figured out a way. Please ignore.

Is it just me or is this repo missing several files and folders? I don't see a trained_models directory for any of the datasets.

Was that a mistake or are the authors holding off on releasing the full repo until after NeurIPS?

Hi!
I have the following problem running your code:
May I ask why the label in the model input of climate experiment is all 0?What does it mean?
Thank you very much!

In the MIMIC pre-processing, a saps2.csv is attempted to be read, but this is not a part of the MIMIC-III dataset. Has this changed since the publication of this repository? Is there a workaround that still allows for reproduction of the results?