exascaleinfolab / bench-vldb20 Goto Github PK

Hi,

I tried to run MRNN algorithm with AIRQ dataset in MCAR setting but I got the following error :

Traceback (most recent call last):
File "", line 1, in
File "testerMRNN.py", line 14, in mrnn_recov
_, Recover_testX = M_RNN(trainZ, trainM, trainT, testZ, testM, testT)
File "M_RNN.py", line 20, in M_RNN
seq_length = len(trainZ[0,:,0])
IndexError: too many indices for array

How can I resolve this issue?

Hi,

The README file for the repo says that the codebase supports BRITS implementation, though I was unable to locate the files for the same. Also in the arguments section, --alg doesn't seem to support brits flag. Clarity on this would be appreciated

Parikshit

Hi,

Is there a way to export the datasets with the generated missing values into files? What command/function should I call if I just want to add the missing data, but not running the algorithms? I want to take them into my own framework to compare my results with those of the benchmark.

Thanks!!

Hello, this is not at issue at all, I just wanted to know if is possible that you could share the blackout MAE and RMSE results without need to run the experiments? I will appreciate it. Regards